diff --git a/.gitignore b/.gitignore index 1de163bd..dbb32598 100644 --- a/.gitignore +++ b/.gitignore @@ -26,4 +26,10 @@ GRTAGS GPATH GTAGS docs/book +test_sql #*# + +# commitizen +node_modules +package.json +package-lock.json diff --git a/.gitmodules b/.gitmodules index e88278f8..092668f6 100644 --- a/.gitmodules +++ b/.gitmodules @@ -10,3 +10,6 @@ [submodule "deps/3rd/benchmark"] path = deps/3rd/benchmark url = https://github.com/google/benchmark +[submodule "deps/3rd/miniob"] + path = deps/3rd/miniob + url = https://github.com/oceanbase/miniob diff --git a/.vscode/launch.json b/.vscode/launch.json deleted file mode 100644 index 18d82e74..00000000 --- a/.vscode/launch.json +++ /dev/null @@ -1,29 +0,0 @@ -{ - // Use IntelliSense to learn about possible attributes. - // Hover to view descriptions of existing attributes. - // For more information, visit: https://go.microsoft.com/fwlink/?linkid=830387 - "version": "0.2.0", - "configurations": [ - { - "type": "cppdbg", - "request": "launch", - "name": "Debug", - "program": "${workspaceFolder}/${defaultBuildTask}/bin/observer", - "args": ["-f", "${workspaceFolder}/etc/observer.ini", "-P", "cli"], - "cwd": "${workspaceFolder}/${defaultBuildTask}/", - "internalConsoleOptions": "openOnSessionStart", - "osx": { - "MIMode": "lldb", - "externalConsole":true - } - }, - { - "type": "lldb", - "request": "launch", - "name": "LLDB", - "program": "${workspaceFolder}/${defaultBuildTask}/bin/observer", - "args": ["-f", "${workspaceFolder}/etc/observer.ini", "-P", "cli"], - "cwd": "${workspaceFolder}/${defaultBuildTask}/" - } - ] -} \ No newline at end of file diff --git a/.vscode/tasks.json b/.vscode/tasks.json deleted file mode 100644 index c64205fa..00000000 --- a/.vscode/tasks.json +++ /dev/null @@ -1,32 +0,0 @@ -{ - // See https://go.microsoft.com/fwlink/?LinkId=733558 - // for the documentation about the tasks.json format - "version": "2.0.0", - "tasks": [ - { - "label": "init", - "type": "shell", - "command": "sudo -E env PATH=$PATH bash ${workspaceFolder}/build.sh init" - }, - { - "label": "build_debug", - "type": "shell", - "command": "bash build.sh debug", - "problemMatcher": [], - "group": { - "kind": "build", - "isDefault": true - } - }, - { - "label": "build_release", - "type": "shell", - "command": "bash build.sh release" - }, - { - "label": "gen_parser", - "type": "shell", - "command": "cd ${workspaceFolder}/src/observer/sql/parser && bash gen_parser.sh" - } - ] -} diff --git a/CMakeLists.txt b/CMakeLists.txt index 31b5917f..312e8b57 100644 --- a/CMakeLists.txt +++ b/CMakeLists.txt @@ -26,22 +26,22 @@ OPTION(USE_MUSL_LIBC "Use musl libc" OFF) MESSAGE(STATUS "HOME dir: $ENV{HOME}") #SET(ENV{变量名} 值) -IF(WIN32) +IF (WIN32) MESSAGE(STATUS "This is windows.") ADD_DEFINITIONS(-DWIN32) -ELSEIF(WIN64) +ELSEIF (WIN64) MESSAGE(STATUS "This is windows.") ADD_DEFINITIONS(-DWIN64) -ELSEIF(APPLE) +ELSEIF (APPLE) MESSAGE(STATUS "This is apple") # normally __MACH__ has already been defined - ADD_DEFINITIONS(-D__MACH__ ) -ELSEIF(UNIX) + ADD_DEFINITIONS(-D__MACH__) +ELSEIF (UNIX) MESSAGE(STATUS "This is UNIX") ADD_DEFINITIONS(-DUNIX -DLINUX) -ELSE() +ELSE () MESSAGE(STATUS "This is UNKNOW OS") -ENDIF(WIN32) +ENDIF (WIN32) # This is for clangd plugin for vscode # mute sign-compare error in lex/yacc @@ -52,23 +52,39 @@ IF (ENABLE_NOPIE) ENDIF (ENABLE_NOPIE) # Requires support for avx2 -IF(USE_SIMD) +IF (USE_SIMD) SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -mavx2") ADD_DEFINITIONS(-DUSE_SIMD) -ENDIF(USE_SIMD) +ENDIF (USE_SIMD) -IF(DEBUG) - MESSAGE(STATUS "DEBUG has been set as TRUE ${DEBUG}") - SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -O0 -g -DDEBUG ") +# Set the default build type to Release +IF (NOT CMAKE_BUILD_TYPE) + SET(CMAKE_BUILD_TYPE "Release" CACHE STRING "BUILD TYPE" FORCE) +ENDIF () + +# Convert CMAKE_BUILD_TYPE to lowercase for case-insensitive comparison +STRING(TOLOWER "${CMAKE_BUILD_TYPE}" CMAKE_BUILD_TYPE) + +# Configure compilation options based on the build type +IF (CMAKE_BUILD_TYPE STREQUAL "debug") + MESSAGE(STATUS "Building in Debug mode") + SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -O0 -g -DDEBUG") ADD_DEFINITIONS(-DENABLE_DEBUG) -ELSEIF(NOT DEFINED ENV{DEBUG}) - MESSAGE(STATUS "Disable debug") - SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -O2 -g ") -ELSE() - MESSAGE(STATUS "Enable debug") - SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -O0 -g -DDEBUG") +ELSEIF (CMAKE_BUILD_TYPE STREQUAL "release") + MESSAGE(STATUS "Building in Release mode") + SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -O2") +ELSEIF (CMAKE_BUILD_TYPE STREQUAL "relwithdebinfo") + MESSAGE(STATUS "Building in RelWithDebInfo mode") + SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -O2 -g -DDEBUG") ADD_DEFINITIONS(-DENABLE_DEBUG) -ENDIF(DEBUG) +ELSEIF (CMAKE_BUILD_TYPE STREQUAL "minsizerel") + MESSAGE(STATUS "Building in MinSizeRel mode") + SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -Os") + ADD_DEFINITIONS(-DENABLE_MIN_SIZE) +ELSE () + MESSAGE(WARNING "Unknown build type: ${CMAKE_BUILD_TYPE}. Defaulting to Release.") + SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -O2") +ENDIF () IF (CONCURRENCY) MESSAGE(STATUS "CONCURRENCY is ON") @@ -79,7 +95,7 @@ ENDIF (CONCURRENCY) MESSAGE(STATUS "CMAKE_CXX_COMPILER_ID is " ${CMAKE_CXX_COMPILER_ID}) IF ("${CMAKE_CXX_COMPILER_ID}" STREQUAL "GNU" AND ${STATIC_STDLIB}) ADD_LINK_OPTIONS(-static-libgcc -static-libstdc++) -ENDIF() +ENDIF () IF(USE_MUSL_LIBC) ADD_DEFINITIONS(-D__MUSL__) @@ -95,8 +111,8 @@ IF (ENABLE_ASAN) SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -fno-omit-frame-pointer -fsanitize=address -fsanitize-address-use-after-scope") IF ("${CMAKE_CXX_COMPILER_ID}" STREQUAL "GNU" AND ${STATIC_STDLIB}) ADD_LINK_OPTIONS(-static-libasan) - ENDIF() -ENDIF() + ENDIF () +ENDIF () IF (ENABLE_TSAN) # supported flags @@ -110,7 +126,7 @@ IF (ENABLE_TSAN) # -Qunused-arguments 有些编译器不支持，所以先删掉 IF ("${CMAKE_CXX_COMPILER_ID}" STREQUAL "GNU" AND ${STATIC_STDLIB}) ADD_LINK_OPTIONS(-static-libtsan) - ENDIF() + ENDIF () SET(CMAKE_SHARED_LINKER_FLAGS "${CMAKE_SHARED_LINKER_FLAGS} ${TSAN_FLAGS}") SET(CMAKE_EXE_LINKER_FLAGS "${CMAKE_EXE_LINKER_FLAGS} ${TSAN_FLAGS}") ENDIF (ENABLE_TSAN) @@ -131,12 +147,12 @@ IF (ENABLE_UBSAN) ENDIF (ENABLE_UBSAN) IF (CMAKE_INSTALL_PREFIX) - MESSAGE(STATUS "CMAKE_INSTALL_PREFIX has been set as " ${CMAKE_INSTALL_PREFIX} ) -ELSEIF(DEFINED ENV{CMAKE_INSTALL_PREFIX}) + MESSAGE(STATUS "CMAKE_INSTALL_PREFIX has been set as " ${CMAKE_INSTALL_PREFIX}) +ELSEIF (DEFINED ENV{CMAKE_INSTALL_PREFIX}) SET(CMAKE_INSTALL_PREFIX $ENV{CMAKE_INSTALL_PREFIX}) -ELSE() +ELSE () SET(CMAKE_INSTALL_PREFIX /tmp/${PROJECT_NAME}) -ENDIF() +ENDIF () MESSAGE(STATUS "Install target dir is " ${CMAKE_INSTALL_PREFIX}) IF (DEFINED ENV{LD_LIBRARY_PATH}) @@ -147,20 +163,20 @@ IF (DEFINED ENV{LD_LIBRARY_PATH}) ENDIF () IF (EXISTS /usr/local/lib) - LINK_DIRECTORIES (/usr/local/lib) + LINK_DIRECTORIES(/usr/local/lib) ENDIF () IF (EXISTS /usr/local/lib64) - LINK_DIRECTORIES (/usr/local/lib64) + LINK_DIRECTORIES(/usr/local/lib64) ENDIF () INCLUDE_DIRECTORIES(. ${PROJECT_SOURCE_DIR}/deps /usr/local/include) -IF(WITH_UNIT_TESTS) +IF (WITH_UNIT_TESTS) IF (ENABLE_COVERAGE) SET(CMAKE_COMMON_FLAGS "${CMAKE_COMMON_FLAGS} -fprofile-arcs -ftest-coverage") ENDIF (ENABLE_COVERAGE) enable_testing() -ENDIF(WITH_UNIT_TESTS) +ENDIF (WITH_UNIT_TESTS) SET(CMAKE_CXX_FLAGS ${CMAKE_COMMON_FLAGS}) SET(CMAKE_C_FLAGS ${CMAKE_COMMON_FLAGS}) @@ -175,9 +191,8 @@ ADD_SUBDIRECTORY(tools) IF (WITH_BENCHMARK) ADD_SUBDIRECTORY(benchmark) -ENDIF(WITH_BENCHMARK) +ENDIF (WITH_BENCHMARK) -IF(WITH_UNIT_TESTS) +IF (WITH_UNIT_TESTS) ADD_SUBDIRECTORY(unittest) -ENDIF(WITH_UNIT_TESTS) - +ENDIF (WITH_UNIT_TESTS) diff --git a/benchmark/aggregate_hash_table_performance_test.cpp b/benchmark/aggregate_hash_table_performance_test.cpp index c2040bb5..20be695e 100644 --- a/benchmark/aggregate_hash_table_performance_test.cpp +++ b/benchmark/aggregate_hash_table_performance_test.cpp @@ -48,8 +48,8 @@ class DISABLED_StandardAggregateHashTableBenchmark : public AggregateHashTableBe { AggregateHashTableBenchmark::SetUp(state); - AggregateExpr aggregate_expr(AggregateExpr::Type::SUM, nullptr); - vector aggregate_exprs; + AggregateFunctionExpr aggregate_expr(AggregateFunctionExpr::Type::SUM, nullptr); + vector aggregate_exprs; aggregate_exprs.push_back(&aggregate_expr); standard_hash_table_ = make_unique(aggregate_exprs); } @@ -75,7 +75,7 @@ class DISABLED_LinearProbingAggregateHashTableBenchmark : public AggregateHashTa { AggregateHashTableBenchmark::SetUp(state); - linear_probing_hash_table_ = make_unique>(AggregateExpr::Type::SUM); + linear_probing_hash_table_ = make_unique>(AggregateFunctionExpr::Type::SUM); } protected: diff --git a/build.sh b/build.sh index 0a914d67..8e036ce3 100755 --- a/build.sh +++ b/build.sh @@ -18,21 +18,51 @@ function usage { echo "Usage:" echo "./build.sh -h" - echo "./build.sh init # install dependence" - echo "./build.sh clean" + echo "./build.sh init # Initialize and install dependencies" + echo "./build.sh clean # Clean up build directories" + echo "./build.sh gen_parser # Generate parser files" + echo "./build.sh style # Apply coding style to source files" + echo "./build.sh test [test_case] # Run tests, optionally specify a test case" echo "./build.sh [BuildType] [--make [MakeOptions]]" echo "" echo "OPTIONS:" - echo "BuildType => debug(default), release" - echo "MakeOptions => Options to make command, default: -j N" + echo " -h Show this help message" + echo " init Initialize and install dependencies" + echo " clean Clean up all build directories" + echo " gen_parser Generate parser files in src/observer/sql/parser" + echo " style Apply coding style to source files in ./src" + echo " test [test_case] Run tests, optionally specify a test case to run" + echo " [BuildType] Specify build type: debug (default), release, relwithdebinfo, minsizerel" + echo " --make [MakeOptions] Options to pass to the make command, default: -jN" echo "" echo "Examples:" - echo "# Init." + echo "# Show help." + echo "./build.sh -h" + echo "" + echo "# Initialize the project and install dependencies." echo "./build.sh init" echo "" - echo "# Build by debug mode and make with -j24." + echo "# Clean up all build directories." + echo "./build.sh clean" + echo "" + echo "# Generate parser files." + echo "./build.sh gen_parser" + echo "" + echo "# Apply coding style to source files." + echo "./build.sh style" + echo "" + echo "# Run all tests." + echo "./build.sh test" + echo "" + echo "# Run a specific test case." + echo "./build.sh test my_test_case" + echo "" + echo "# Build in debug mode (default) and make with -j24." echo "./build.sh debug --make -j24" + echo "" + echo "# Build in release mode." + echo "./build.sh release" } function parse_args @@ -150,12 +180,21 @@ function do_clean function build { set -- "${BUILD_ARGS[@]}" - case "x$1" in + # 将传递的所有参数转换为小写 + BUILD_TYPE=$(echo "$1" | tr '[:upper:]' '[:lower:]') + + case "x$BUILD_TYPE" in xrelease) - do_build "$@" -DCMAKE_BUILD_TYPE=RelWithDebInfo -DDEBUG=OFF + do_build "$@" -DCMAKE_BUILD_TYPE=Release + ;; + xrelwithdebinfo) + do_build "$@" -DCMAKE_BUILD_TYPE=RelWithDebInfo ;; xdebug) - do_build "$@" -DCMAKE_BUILD_TYPE=Debug -DDEBUG=ON + do_build "$@" -DCMAKE_BUILD_TYPE=Debug + ;; + xminsizerel) + do_build "$@" -DCMAKE_BUILD_TYPE=MinSizeRel ;; *) BUILD_ARGS=(debug "${BUILD_ARGS[@]}") @@ -164,6 +203,49 @@ function build esac } +function gen_parser +{ + echo "generate parser..." + cd ${TOPDIR}/src/observer/sql/parser + ./gen_parser.sh + echo "generate parser done" + cd ${TOPDIR} +} + +function style +{ + # Check if .clang-format file exists + if [ ! -f .clang-format ]; then + echo "Error: .clang-format file not found in the current directory." + exit 1 + fi + + # 设置要格式化的文件扩展名 + EXTENSIONS=("c" "h" "cpp" "hpp") + + # 查找并格式化文件 + for ext in "${EXTENSIONS[@]}"; do + find ./src -type f -name "*.$ext" -exec clang-format -i {} + + done + + echo "Formatting complete!" +} + +function run_tests { + echo "Running test(s)..." + cd ${TOPDIR}/test/case || exit + + if [[ -z "$1" ]]; then + # 如果没有提供测试用例名，直接执行 + python3 miniob_test.py + else + # 如果提供了测试用例名，执行带参数的命令 + TEST_CASE="$1" + echo "Running test case: $TEST_CASE" + python3 miniob_test.py --test-cases="$TEST_CASE" + fi +} + function main { case "$1" in @@ -179,6 +261,15 @@ function main clean) do_clean ;; + gen_parser) + gen_parser + ;; + style) + style + ;; + test) + run_tests "$2" + ;; *) parse_args build diff --git a/deps/3rd/benchmark b/deps/3rd/benchmark index f7547e29..761305ec 160000 --- a/deps/3rd/benchmark +++ b/deps/3rd/benchmark @@ -1 +1 @@ -Subproject commit f7547e29ccaed7b64ef4f7495ecfff1c9f6f3d03 +Subproject commit 761305ec3b33abf30e08d50eb829e19a802581cc diff --git a/deps/3rd/googletest b/deps/3rd/googletest index 974e18ee..df1544bc 160000 --- a/deps/3rd/googletest +++ b/deps/3rd/googletest @@ -1 +1 @@ -Subproject commit 974e18ee6f146a2418f9cea83170c640e7d622d6 +Subproject commit df1544bcee0c7ce35cd5ea0b3eb8cc81855a4140 diff --git a/deps/3rd/jsoncpp b/deps/3rd/jsoncpp index 8190e061..bd25fc5e 160000 --- a/deps/3rd/jsoncpp +++ b/deps/3rd/jsoncpp @@ -1 +1 @@ -Subproject commit 8190e061bc2d95da37479a638aa2c9e483e58ec6 +Subproject commit bd25fc5ea0e14d19e1451632205c8b99ec0b1c09 diff --git a/deps/3rd/libevent b/deps/3rd/libevent index 5df3037d..f6e426c2 160000 --- a/deps/3rd/libevent +++ b/deps/3rd/libevent @@ -1 +1 @@ -Subproject commit 5df3037d10556bfcb675bc73e516978b75fc7bc7 +Subproject commit f6e426c299acd80e85fef24c426fea45d36ace8d diff --git a/deps/3rd/miniob b/deps/3rd/miniob new file mode 160000 index 00000000..6f7d7406 --- /dev/null +++ b/deps/3rd/miniob @@ -0,0 +1 @@ +Subproject commit 6f7d74063db39b5bab5a6f9378f83536a5009e98 diff --git a/deps/common/lang/defer.h b/deps/common/lang/defer.h index 8035f9d1..a0dcbe0e 100644 --- a/deps/common/lang/defer.h +++ b/deps/common/lang/defer.h @@ -40,4 +40,4 @@ class DeferHelper #define SCOPE_UNIQUE_NAME(B) _SCOPE_UNIQUE_NAME(B, __LINE__) -#define DEFER(...) common::DeferHelper SCOPE_UNIQUE_NAME(defer_helper_)([&]() { __VA_ARGS__; }) +#define DEFER(...) common::DeferHelper SCOPE_UNIQUE_NAME(defer_helper_)([&]() { __VA_ARGS__; }) \ No newline at end of file diff --git a/deps/common/log/log.h b/deps/common/log/log.h index f51b798c..7f1cd868 100644 --- a/deps/common/log/log.h +++ b/deps/common/log/log.h @@ -58,8 +58,8 @@ typedef enum class Log { public: - Log(const string &log_name, const LOG_LEVEL log_level = LOG_LEVEL_INFO, - const LOG_LEVEL console_level = LOG_LEVEL_WARN); + Log(const string &log_name, const LOG_LEVEL log_level = LOG_LEVEL_INFO, + const LOG_LEVEL console_level = LOG_LEVEL_WARN); ~Log(void); static int init(const string &log_file); @@ -162,7 +162,7 @@ class Log class LoggerFactory { public: - LoggerFactory(); + LoggerFactory(); virtual ~LoggerFactory(); static int init(const string &log_file, Log **logger, LOG_LEVEL log_level = LOG_LEVEL_INFO, @@ -228,7 +228,12 @@ extern Log *g_log; #define LOG_DEFAULT(fmt, ...) LOG_OUTPUT(common::g_log->get_log_level(), fmt, ##__VA_ARGS__) #define LOG_PANIC(fmt, ...) LOG_OUTPUT(common::LOG_LEVEL_PANIC, fmt, ##__VA_ARGS__) #define LOG_ERROR(fmt, ...) LOG_OUTPUT(common::LOG_LEVEL_ERR, fmt, ##__VA_ARGS__) -#define LOG_WARN(fmt, ...) LOG_OUTPUT(common::LOG_LEVEL_WARN, fmt, ##__VA_ARGS__) +#define LOG_WARN(fmt, ...) \ + do { \ + assert((fmt != nullptr) && "LOG_WARN: format string must not be null"); \ + LOG_OUTPUT(common::LOG_LEVEL_WARN, fmt, ##__VA_ARGS__); \ + } while (0) + #define LOG_INFO(fmt, ...) LOG_OUTPUT(common::LOG_LEVEL_INFO, fmt, ##__VA_ARGS__) #define LOG_DEBUG(fmt, ...) LOG_OUTPUT(common::LOG_LEVEL_DEBUG, fmt, ##__VA_ARGS__) #define LOG_TRACE(fmt, ...) LOG_OUTPUT(common::LOG_LEVEL_TRACE, fmt, ##__VA_ARGS__) @@ -315,12 +320,12 @@ int Log::out(const LOG_LEVEL console_level, const LOG_LEVEL log_level, T &msg) #ifndef ASSERT #ifdef DEBUG -#define ASSERT(expression, description, ...) \ - do { \ - if (!(expression)) { \ - LOG_PANIC(description, ##__VA_ARGS__); \ - assert(expression); \ - } \ +#define ASSERT(expression, description, ...) \ + do { \ + if (!(expression)) { \ + LOG_PANIC(description, ## __VA_ARGS__); \ + assert(expression); \ + } \ } while (0) #else // DEBUG @@ -334,7 +339,7 @@ int Log::out(const LOG_LEVEL console_level, const LOG_LEVEL log_level, T &msg) #ifndef TRACE #ifdef DEBUG -#define TRACE(format, ...) LOG_TRACE(format, ##__VA_ARGS__) +#define TRACE(format, ...) LOG_TRACE(format, ## __VA_ARGS__) #else // DEBUG #define TRACE(...) #endif // DEBUG diff --git a/docs/docs/game/miniob-vectordb.md b/docs/docs/game/miniob-vectordb.md index 52c93e3c..105204f5 100644 --- a/docs/docs/game/miniob-vectordb.md +++ b/docs/docs/game/miniob-vectordb.md @@ -60,7 +60,7 @@ select embedding + '[1.5,2.3,3.3]', embedding - '[1,2,3]', '[1,2,3]' - embedding * 语法：inner_product(vector A, vector B) * 计算公式：$[ D = \mathbf{A} \cdot \mathbf{B} = a_1 b_1 + a_2 b_2 + ... + a_n b_n = \sum_{i=1}^{n} a_i b_i ]$ - * 距离表达式的计算精度为保留2位小数。 + * 距离表达式的计算精度为保留2位小数。 ### 题目二：向量类型扩展 @@ -88,7 +88,7 @@ IVF-Flat 是一种常见的高效近似最近邻（Approximate Nearest Neighbor ![ivfflat](images/ivfflat.png) 以图像搜索为例的 IVF-Flat 向量索引工作流程如下： 1. 首先使用 K-Means 算法将所有图像的特征向量分成 1000 个簇。每个簇表示一部分相似的图片。例如，簇 1 可能包含大部分是风景图像的向量，簇 2 可能包含大部分是人物图像的向量，以此类推。对于每个簇，向量索引都维护一个倒排文件索引，存储属于这个簇的所有图像的索引。 -比如，簇 1 对应的倒排文件索引可能包括 [image_23, image_45, image_78...]，这些是属于簇 1 的图像的索引。 + 比如，簇 1 对应的倒排文件索引可能包括 [image_23, image_45, image_78...]，这些是属于簇 1 的图像的索引。 1. 查询过程：用户指定一张新图像进行搜索。首先，将这张图片通过同样的特征提取模型转换成一个高维向量。然后，通过计算这个向量与1000个质心的距离，找到离它最近的几个质心（假设是前5个）。在这几个质心对应的簇中，你进一步进行精确的向量搜索，找到最相似的图像。 @@ -166,7 +166,7 @@ pip install -r requirements.txt /root/miniob/build_release/bin/observer -s /tmp/miniob.sock -P mysql ``` 4. 运行 ann-benchmark. -注意：需要将 `algorithms/miniob/config.yml` 中的 `arg_groups: [{unix_socket: "/tmp/miniob.sock"}]` 修改为 miniob 实际使用的 unix socket 文件地址 + 注意：需要将 `algorithms/miniob/config.yml` 中的 `arg_groups: [{unix_socket: "/tmp/miniob.sock"}]` 修改为 miniob 实际使用的 unix socket 文件地址 ```bash # 示例命令 python3 run.py --dataset fashion-mnist-784-euclidean --docker-tag ann-benchmarks-miniob --local --timeout 100 --runs 1 diff --git a/src/obclient/client.cpp b/src/obclient/client.cpp index 66c2f7e4..1808cbab 100644 --- a/src/obclient/client.cpp +++ b/src/obclient/client.cpp @@ -8,9 +8,7 @@ EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. See the Mulan PSL v2 for more details. */ -// -// Created by Longda on 2021 -// +#if 1 #include #include @@ -35,7 +33,7 @@ See the Mulan PSL v2 for more details. */ #include "readline/readline.h" #endif -#define MAX_MEM_BUFFER_SIZE 8192 +#define MAX_MEM_BUFFER_SIZE 131072 #define PORT_DEFAULT 6789 using namespace std; @@ -245,3 +243,5 @@ int main(int argc, char *argv[]) return 0; } + +#endif \ No newline at end of file diff --git a/src/obclient/client_order_by.cpp b/src/obclient/client_order_by.cpp new file mode 100644 index 00000000..f13df4e3 --- /dev/null +++ b/src/obclient/client_order_by.cpp @@ -0,0 +1,277 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +#if 0 + +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include +#include + +#include "common/defs.h" +#include "common/lang/string.h" + +#ifdef USE_READLINE +#include "readline/history.h" +#include "readline/readline.h" +#endif + +const int TABLE_RECORD_NUMBER = 20; + +#define MAX_MEM_BUFFER_SIZE 131072 +#define PORT_DEFAULT 6789 + +using namespace std; +using namespace common; + +#ifdef USE_READLINE +const string HISTORY_FILE = string(getenv("HOME")) + "/.miniob.history"; +time_t last_history_write_time = 0; + +char *my_readline(const char *prompt) +{ + int size = history_length; + if (size == 0) { + read_history(HISTORY_FILE.c_str()); + + FILE *fp = fopen(HISTORY_FILE.c_str(), "a"); + if (fp != nullptr) { + fclose(fp); + } + } + + char *line = readline(prompt); + if (line != nullptr && line[0] != 0) { + add_history(line); + if (time(NULL) - last_history_write_time > 5) { + write_history(HISTORY_FILE.c_str()); + } + } + return line; +} +#else // USE_READLINE +char *my_readline(const char *prompt) +{ + char *buffer = (char *)malloc(MAX_MEM_BUFFER_SIZE); + if (nullptr == buffer) { + fprintf(stderr, "failed to alloc line buffer"); + return nullptr; + } + fprintf(stdout, "%s", prompt); + char *s = fgets(buffer, MAX_MEM_BUFFER_SIZE, stdin); + if (nullptr == s) { + fprintf(stderr, "failed to read message from console"); + free(buffer); + return nullptr; + } + return buffer; +} +#endif // USE_READLINE + +bool is_exit_command(const char *cmd) +{ + return 0 == strncasecmp("exit", cmd, 4) || 0 == strncasecmp("bye", cmd, 3) || 0 == strncasecmp("\\q", cmd, 2); +} + +int init_unix_sock(const char *unix_sock_path) +{ + int sockfd = socket(PF_UNIX, SOCK_STREAM, 0); + if (sockfd < 0) { + fprintf(stderr, "failed to create unix socket. %s", strerror(errno)); + return -1; + } + + struct sockaddr_un sockaddr; + memset(&sockaddr, 0, sizeof(sockaddr)); + sockaddr.sun_family = PF_UNIX; + snprintf(sockaddr.sun_path, sizeof(sockaddr.sun_path), "%s", unix_sock_path); + + if (connect(sockfd, (struct sockaddr *)&sockaddr, sizeof(sockaddr)) < 0) { + fprintf(stderr, "failed to connect to server. unix socket path '%s'. error %s", sockaddr.sun_path, strerror(errno)); + close(sockfd); + return -1; + } + return sockfd; +} + +int init_tcp_sock(const char *server_host, int server_port) +{ + struct hostent *host; + struct sockaddr_in serv_addr; + + if ((host = gethostbyname(server_host)) == NULL) { + fprintf(stderr, "gethostbyname failed. errmsg=%d:%s\n", errno, strerror(errno)); + return -1; + } + + int sockfd; + if ((sockfd = socket(AF_INET, SOCK_STREAM, 0)) == -1) { + fprintf(stderr, "create socket error. errmsg=%d:%s\n", errno, strerror(errno)); + return -1; + } + + serv_addr.sin_family = AF_INET; + serv_addr.sin_port = htons(server_port); + serv_addr.sin_addr = *((struct in_addr *)host->h_addr); + bzero(&(serv_addr.sin_zero), 8); + + if (connect(sockfd, (struct sockaddr *)&serv_addr, sizeof(struct sockaddr)) == -1) { + fprintf(stderr, "Failed to connect. errmsg=%d:%s\n", errno, strerror(errno)); + close(sockfd); + return -1; + } + return sockfd; +} + +const char *startup_tips = R"( +Welcome to the OceanBase database implementation course. + +Copyright (c) 2021 OceanBase and/or its affiliates. + +Learn more about OceanBase at https://github.com/oceanbase/oceanbase +Learn more about MiniOB at https://github.com/oceanbase/miniob + +)"; + +// Function to send SQL query and receive response +void send_sql(int sockfd, const char *sql) +{ + char send_buf[MAX_MEM_BUFFER_SIZE]; + printf("Sending SQL: %s\n", sql); + + if (write(sockfd, sql, strlen(sql) + 1) == -1) { + fprintf(stderr, "send error: %d:%s \n", errno, strerror(errno)); + exit(1); + } + + memset(send_buf, 0, sizeof(send_buf)); + int len = 0; + while ((len = recv(sockfd, send_buf, MAX_MEM_BUFFER_SIZE, 0)) > 0) { + bool msg_end = false; + for (int i = 0; i < len; i++) { + if (0 == send_buf[i]) { + msg_end = true; + break; + } + printf("%c", send_buf[i]); + } + if (msg_end) { + break; + } + memset(send_buf, 0, MAX_MEM_BUFFER_SIZE); + } + + if (len < 0) { + fprintf(stderr, "Connection was broken: %s\n", strerror(errno)); + exit(1); + } else if (len == 0) { + printf("Connection closed by server.\n"); + exit(1); + } +} + +int main(int argc, char *argv[]) +{ + printf("%s", startup_tips); + + const char *unix_socket_path = nullptr; + const char *server_host = "127.0.0.1"; + int server_port = PORT_DEFAULT; + int opt; + extern char *optarg; + while ((opt = getopt(argc, argv, "s:h:p:")) > 0) { + switch (opt) { + case 's': unix_socket_path = optarg; break; + case 'p': server_port = atoi(optarg); break; + case 'h': server_host = optarg; break; + } + } + + int sockfd; + + if (unix_socket_path != nullptr) { + sockfd = init_unix_sock(unix_socket_path); + } else { + sockfd = init_tcp_sock(server_host, server_port); + } + if (sockfd < 0) { + return 1; + } + + // Send CREATE TABLE statements + send_sql(sockfd, "CREATE TABLE big_order_by_0 (id INT, addr CHAR(100), num INT, price FLOAT, birthday DATE);"); + send_sql(sockfd, "CREATE TABLE big_order_by_1 (id INT, addr CHAR(100), num INT, price FLOAT, birthday DATE);"); + send_sql(sockfd, "CREATE TABLE big_order_by_2 (id INT, addr CHAR(100), num INT, price FLOAT, birthday DATE);"); + send_sql(sockfd, "CREATE TABLE big_order_by_3 (id INT, addr CHAR(100), num INT, price FLOAT, birthday DATE);"); + + // Insert data (for simplicity, inserting a small amount of random data) + for (int i = 1; i <= TABLE_RECORD_NUMBER; i++) { + char insert_sql[512]; + snprintf(insert_sql, + sizeof(insert_sql), + "INSERT INTO big_order_by_0 VALUES (%d, 'addr%d', %d, %.2f, '2022-01-01');", + i, + i, + rand() % 1000, + (float)(rand() % 10000) / 100); + send_sql(sockfd, insert_sql); + + snprintf(insert_sql, + sizeof(insert_sql), + "INSERT INTO big_order_by_1 VALUES (%d, 'addr%d', %d, %.2f, '2022-01-01');", + i, + i, + rand() % 1000, + (float)(rand() % 10000) / 100); + send_sql(sockfd, insert_sql); + + snprintf(insert_sql, + sizeof(insert_sql), + "INSERT INTO big_order_by_2 VALUES (%d, 'addr%d', %d, %.2f, '2022-01-01');", + i, + i, + rand() % 1000, + (float)(rand() % 10000) / 100); + send_sql(sockfd, insert_sql); + + snprintf(insert_sql, + sizeof(insert_sql), + "INSERT INTO big_order_by_3 VALUES (%d, 'addr%d', %d, %.2f, '2022-01-01');", + i, + i, + rand() % 1000, + (float)(rand() % 10000) / 100); + send_sql(sockfd, insert_sql); + } + + // Send the ORDER BY query + const char *order_by_sql = "SELECT * FROM big_order_by_0, big_order_by_1, big_order_by_2, big_order_by_3 " + "ORDER BY big_order_by_0.addr, big_order_by_2.num, big_order_by_0.price, " + "big_order_by_3.id, big_order_by_1.id, big_order_by_1.num, big_order_by_0.id, " + "big_order_by_0.birthday LIMIT 1;"; + send_sql(sockfd, order_by_sql); + + // Close the socket + close(sockfd); + + return 0; +} + +#endif // #if 0 \ No newline at end of file diff --git a/src/observer/common/ini_setting.h b/src/observer/common/ini_setting.h index 74a35de3..b26eb076 100644 --- a/src/observer/common/ini_setting.h +++ b/src/observer/common/ini_setting.h @@ -22,6 +22,6 @@ See the Mulan PSL v2 for more details. */ #define PORT "PORT" #define PORT_DEFAULT 6789 -#define SOCKET_BUFFER_SIZE 8192 +#define SOCKET_BUFFER_SIZE 131072 #define SESSION_STAGE_NAME "SessionStage" diff --git a/src/observer/common/init.cpp b/src/observer/common/init.cpp index 94d0df96..cbea3120 100644 --- a/src/observer/common/init.cpp +++ b/src/observer/common/init.cpp @@ -129,10 +129,7 @@ void cleanup_log() } } -int prepare_init_seda() -{ - return 0; -} +int prepare_init_seda() { return 0; } int init_global_objects(ProcessParam *process_param, Ini &properties) { @@ -140,9 +137,8 @@ int init_global_objects(ProcessParam *process_param, Ini &properties) int ret = 0; - RC rc = GCTX.handler_->init("miniob", - process_param->trx_kit_name().c_str(), - process_param->durability_mode().c_str()); + RC rc = + GCTX.handler_->init("miniob", process_param->trx_kit_name().c_str(), process_param->durability_mode().c_str()); if (OB_FAIL(rc)) { LOG_ERROR("failed to init handler. rc=%s", strrc(rc)); return -1; diff --git a/src/observer/common/rc.h b/src/observer/common/rc.h index d465c7cf..c9d05af3 100644 --- a/src/observer/common/rc.h +++ b/src/observer/common/rc.h @@ -18,65 +18,86 @@ See the Mulan PSL v2 for more details. */ * @brief 这个文件定义函数返回码/错误码(Return Code) * @enum RC */ - -#define DEFINE_RCS \ - DEFINE_RC(SUCCESS) \ - DEFINE_RC(INVALID_ARGUMENT) \ - DEFINE_RC(UNIMPLEMENTED) \ - DEFINE_RC(SQL_SYNTAX) \ - DEFINE_RC(INTERNAL) \ - DEFINE_RC(NOMEM) \ - DEFINE_RC(NOTFOUND) \ - DEFINE_RC(EMPTY) \ - DEFINE_RC(FULL) \ - DEFINE_RC(EXIST) \ - DEFINE_RC(NOT_EXIST) \ - DEFINE_RC(BUFFERPOOL_OPEN) \ - DEFINE_RC(BUFFERPOOL_NOBUF) \ - DEFINE_RC(BUFFERPOOL_INVALID_PAGE_NUM) \ - DEFINE_RC(RECORD_OPENNED) \ - DEFINE_RC(RECORD_INVALID_RID) \ - DEFINE_RC(RECORD_INVALID_KEY) \ - DEFINE_RC(RECORD_DUPLICATE_KEY) \ - DEFINE_RC(RECORD_NOMEM) \ - DEFINE_RC(RECORD_EOF) \ - DEFINE_RC(RECORD_NOT_EXIST) \ - DEFINE_RC(RECORD_INVISIBLE) \ - DEFINE_RC(SCHEMA_DB_EXIST) \ - DEFINE_RC(SCHEMA_DB_NOT_EXIST) \ - DEFINE_RC(SCHEMA_DB_NOT_OPENED) \ - DEFINE_RC(SCHEMA_TABLE_NOT_EXIST) \ - DEFINE_RC(SCHEMA_TABLE_EXIST) \ - DEFINE_RC(SCHEMA_FIELD_NOT_EXIST) \ - DEFINE_RC(SCHEMA_FIELD_MISSING) \ - DEFINE_RC(SCHEMA_FIELD_TYPE_MISMATCH) \ - DEFINE_RC(SCHEMA_INDEX_NAME_REPEAT) \ - DEFINE_RC(IOERR_READ) \ - DEFINE_RC(IOERR_WRITE) \ - DEFINE_RC(IOERR_ACCESS) \ - DEFINE_RC(IOERR_OPEN) \ - DEFINE_RC(IOERR_CLOSE) \ - DEFINE_RC(IOERR_SEEK) \ - DEFINE_RC(IOERR_TOO_LONG) \ - DEFINE_RC(IOERR_SYNC) \ - DEFINE_RC(LOCKED_UNLOCK) \ - DEFINE_RC(LOCKED_NEED_WAIT) \ - DEFINE_RC(LOCKED_CONCURRENCY_CONFLICT) \ - DEFINE_RC(FILE_EXIST) \ - DEFINE_RC(FILE_NOT_EXIST) \ - DEFINE_RC(FILE_NAME) \ - DEFINE_RC(FILE_BOUND) \ - DEFINE_RC(FILE_CREATE) \ - DEFINE_RC(FILE_OPEN) \ - DEFINE_RC(FILE_NOT_OPENED) \ - DEFINE_RC(FILE_CLOSE) \ - DEFINE_RC(FILE_REMOVE) \ - DEFINE_RC(VARIABLE_NOT_EXISTS) \ - DEFINE_RC(VARIABLE_NOT_VALID) \ - DEFINE_RC(LOGBUF_FULL) \ - DEFINE_RC(LOG_FILE_FULL) \ - DEFINE_RC(LOG_ENTRY_INVALID) \ - DEFINE_RC(UNSUPPORTED) +#define DEFINE_RCS \ + DEFINE_RC(SUCCESS) \ + DEFINE_RC(INVALID_ALIAS) \ + DEFINE_RC(INVALID_ARGUMENT) \ + DEFINE_RC(INVALID_HAVING_CONDITION) \ + DEFINE_RC(UNIMPLEMENTED) \ + DEFINE_RC(SQL_SYNTAX) \ + DEFINE_RC(INTERNAL) \ + DEFINE_RC(NOMEM) \ + DEFINE_RC(NOTFOUND) \ + DEFINE_RC(NUMBER_DIV_ZERO) \ + DEFINE_RC(EMPTY) \ + DEFINE_RC(ERROR_DATE) \ + DEFINE_RC(FULL) \ + DEFINE_RC(EXIST) \ + DEFINE_RC(NOT_EXIST) \ + DEFINE_RC(BUFFERPOOL_OPEN) \ + DEFINE_RC(BUFFERPOOL_NOBUF) \ + DEFINE_RC(BUFFERPOOL_INVALID_PAGE_NUM) \ + DEFINE_RC(RECORD_OPENNED) \ + DEFINE_RC(RECORD_INVALID_RID) \ + DEFINE_RC(RECORD_INVALID_KEY) \ + DEFINE_RC(RECORD_DUPLICATE_KEY) \ + DEFINE_RC(RECORD_NOMEM) \ + DEFINE_RC(RECORD_EOF) \ + DEFINE_RC(RECORD_NOT_EXIST) \ + DEFINE_RC(RECORD_INVISIBLE) \ + DEFINE_RC(SCHEMA_DB_EXIST) \ + DEFINE_RC(SCHEMA_DB_NOT_EXIST) \ + DEFINE_RC(SCHEMA_DB_NOT_OPENED) \ + DEFINE_RC(SCHEMA_TABLE_NOT_EXIST) \ + DEFINE_RC(SCHEMA_TABLE_EXIST) \ + DEFINE_RC(SCHEMA_FIELD_NOT_EXIST) \ + DEFINE_RC(SCHEMA_FIELD_MISSING) \ + DEFINE_RC(SCHEMA_FIELD_TYPE_MISMATCH) \ + DEFINE_RC(SCHEMA_INDEX_NAME_REPEAT) \ + DEFINE_RC(IOERR_READ) \ + DEFINE_RC(IOERR_WRITE) \ + DEFINE_RC(IOERR_ACCESS) \ + DEFINE_RC(IOERR_OPEN) \ + DEFINE_RC(IOERR_CLOSE) \ + DEFINE_RC(IOERR_SEEK) \ + DEFINE_RC(IOERR_TOO_LONG) \ + DEFINE_RC(IOERR_SYNC) \ + DEFINE_RC(VECTOR_PARSE_ERROR) \ + DEFINE_RC(LOCKED_UNLOCK) \ + DEFINE_RC(LOCKED_NEED_WAIT) \ + DEFINE_RC(LOCKED_CONCURRENCY_CONFLICT) \ + DEFINE_RC(FILE_EXIST) \ + DEFINE_RC(FILE_NOT_EXIST) \ + DEFINE_RC(FILE_NAME) \ + DEFINE_RC(FILE_BOUND) \ + DEFINE_RC(FILE_CREATE) \ + DEFINE_RC(FILE_OPEN) \ + DEFINE_RC(FILE_NOT_OPENED) \ + DEFINE_RC(FILE_CLOSE) \ + DEFINE_RC(FILE_REMOVE) \ + DEFINE_RC(VARIABLE_NOT_EXISTS) \ + DEFINE_RC(VARIABLE_NOT_VALID) \ + DEFINE_RC(LOGBUF_FULL) \ + DEFINE_RC(LOG_FILE_FULL) \ + DEFINE_RC(LOG_ENTRY_INVALID) \ + DEFINE_RC(UNSUPPORTED) \ + DEFINE_RC(VALUE_TOO_LONG) \ + DEFINE_RC(VECTOR_LENGTG_ARE_INCONSISTENT) \ + DEFINE_RC(VECTOR_DIM_MISMATCH) \ + DEFINE_RC(TO_LONG_SUBQUERY_EXPR) \ + DEFINE_RC(NOT_NULLABLE_VALUE) \ + DEFINE_RC(NOT_NULL_AFTER_IS) \ + DEFINE_RC(UNKNOWN_FUNCTION) \ + DEFINE_RC(SUBQUERY_RETURNED_MULTIPLE_ROWS) \ + DEFINE_RC(UNKNOWN_TABLE_TYPE) \ + DEFINE_RC(CREATE_INDEX_ON_NON_TABLE_TYPE) \ + DEFINE_RC(JOIN_VIEW_INSERT_ERROR) \ + DEFINE_RC(JOIN_VIEW_DELETE_ERROR) \ + DEFINE_RC(READ_ONLY_VIEW_INSERT_ERROR) \ + DEFINE_RC(READ_ONLY_VIEW_DELETE_ERROR) \ + DEFINE_RC(READ_ONLY_VIEW_UPDATE_ERROR) \ + DEFINE_RC(EXPRESSION_FIELD_NOT_INSERTABLE) \ + DEFINE_RC(EXPRESSION_FIELD_NOT_UPDATEABLE) enum class RC { diff --git a/src/observer/common/type/attr_type.cpp b/src/observer/common/type/attr_type.cpp index 131342db..a15c6f3c 100644 --- a/src/observer/common/type/attr_type.cpp +++ b/src/observer/common/type/attr_type.cpp @@ -8,11 +8,11 @@ EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. See the Mulan PSL v2 for more details. */ - #include "common/lang/string.h" #include "common/type/attr_type.h" -const char *ATTR_TYPE_NAME[] = {"undefined", "chars", "ints", "floats", "vectors", "booleans"}; +const char *ATTR_TYPE_NAME[] = { + "undefined", "chars", "ints", "floats", "booleans", "dates", "nulls", "texts", "vectors", "lists"}; const char *attr_type_to_string(AttrType type) { diff --git a/src/observer/common/type/attr_type.h b/src/observer/common/type/attr_type.h index d1e56cf0..1002d49b 100644 --- a/src/observer/common/type/attr_type.h +++ b/src/observer/common/type/attr_type.h @@ -20,8 +20,11 @@ enum class AttrType CHARS, ///< 字符串类型 INTS, ///< 整数类型(4字节) FLOATS, ///< 浮点数类型(4字节) - VECTORS, ///< 向量类型 BOOLEANS, ///< boolean类型，当前不是由parser解析出来的，是程序内部使用的 + DATES, ///< 日期类型(4字节) + NULLS, ///< 空字段 + TEXTS, ///< text 超长字段(4096字节) + VECTORS, ///< 向量 MAXTYPE, ///< 请在 UNDEFINED 与 MAXTYPE 之间增加新类型 }; diff --git a/src/observer/common/type/boolean_type.cpp b/src/observer/common/type/boolean_type.cpp new file mode 100644 index 00000000..4417d28a --- /dev/null +++ b/src/observer/common/type/boolean_type.cpp @@ -0,0 +1,22 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +// +// Created by 胡鑫 on 24-10-19. +// + +#include "boolean_type.h" +#include "common/value.h" + +RC BooleanType::to_string(const Value &val, string &result) const +{ + result = val.get_boolean() ? "TRUE" : "FALSE"; + return RC::SUCCESS; +} diff --git a/src/observer/common/type/boolean_type.h b/src/observer/common/type/boolean_type.h new file mode 100644 index 00000000..c039f345 --- /dev/null +++ b/src/observer/common/type/boolean_type.h @@ -0,0 +1,28 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +#pragma once + +#include "common/rc.h" +#include "common/type/data_type.h" + +// +// Created by 胡鑫 on 24-10-19. +// + +class BooleanType : public DataType +{ +public: + BooleanType() : DataType(AttrType::BOOLEANS) {} + + virtual ~BooleanType() = default; + + RC to_string(const Value &val, string &result) const override; +}; diff --git a/src/observer/common/type/char_type.cpp b/src/observer/common/type/char_type.cpp index 6c93114a..4fd57869 100644 --- a/src/observer/common/type/char_type.cpp +++ b/src/observer/common/type/char_type.cpp @@ -8,10 +8,14 @@ EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. See the Mulan PSL v2 for more details. */ +#include + +#include "common/utils.h" #include "common/lang/comparator.h" #include "common/log/log.h" #include "common/type/char_type.h" #include "common/value.h" +#include "common/type/attr_type.h" int CharType::compare(const Value &left, const Value &right) const { @@ -26,9 +30,66 @@ RC CharType::set_value_from_str(Value &val, const string &data) const return RC::SUCCESS; } -RC CharType::cast_to(const Value &val, AttrType type, Value &result) const +RC CharType::cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion) const { switch (type) { + case AttrType::TEXTS: { + if (val.length() > 65535) { + return RC::VALUE_TOO_LONG; + } + result.set_text(val.value_.pointer_value_, val.length_); + } break; + case AttrType::DATES: { + int date_val; + RC rc = parse_date(val.value_.pointer_value_, date_val); + if (rc != RC::SUCCESS) { + return rc; + } + result.set_date(date_val); + } break; + // 字符串转数字 + // 如果字符串刚好是一个数字，则转换为对应的数字（如'1'转换为1，'2.1'转换为2.1） + // 如果字符串的前缀是一个数字，则转换前缀数字部分为数字（如'1a1'转换为1，'2.1a'转换为2.1） + // 如果字符串前缀不是任何合法的数字，则转换为0（不需要考虑前导符号 '+' '-'） + // 如果转换数字溢出怎么处理?（不考虑） + // 是否考虑十六进制/八进制?（不考虑） + case AttrType::INTS: { + // WHERE id < '1.5a'; + // 先当浮点数解析，浮点数可以兼容整型，再四舍五入转 int + float float_val; + RC rc = parse_float_prefix(val.value_.pointer_value_, float_val); + if (rc != RC::SUCCESS) { + return rc; + } + int int_val = static_cast(float_val); + // 是整数 + if (int_val == float_val || !allow_type_promotion) { + result.set_int(int_val); + } else { + // 为浮点数，类型提升 + result.set_float(float_val); + } + } break; + case AttrType::FLOATS: { + float float_val; + RC rc = parse_float_prefix(val.value_.pointer_value_, float_val); + if (rc != RC::SUCCESS) { + return rc; + } + result.set_float(float_val); + } break; + case AttrType::VECTORS: { + float *array = nullptr; + int length = 0; + + // 解析字符串为 float 数组 + RC rc = parse_vector_from_string(val.value_.pointer_value_, array, length); + if (rc != RC::SUCCESS) { + return rc; + } + + result.set_vector(array, length); + } break; default: return RC::UNIMPLEMENTED; } return RC::SUCCESS; @@ -36,9 +97,18 @@ RC CharType::cast_to(const Value &val, AttrType type, Value &result) const int CharType::cast_cost(AttrType type) { - if (type == AttrType::CHARS) { + if (type == AttrType::CHARS || type == AttrType::TEXTS) { return 0; } + if (type == AttrType::DATES) { + return 1; + } + if (type == AttrType::INTS) { + return 1; + } + if (type == AttrType::FLOATS) { + return 1; + } return INT32_MAX; } @@ -48,4 +118,4 @@ RC CharType::to_string(const Value &val, string &result) const ss << val.value_.pointer_value_; result = ss.str(); return RC::SUCCESS; -} \ No newline at end of file +} diff --git a/src/observer/common/type/char_type.h b/src/observer/common/type/char_type.h index a5610143..be49239b 100644 --- a/src/observer/common/type/char_type.h +++ b/src/observer/common/type/char_type.h @@ -26,11 +26,11 @@ class CharType : public DataType int compare(const Value &left, const Value &right) const override; - RC cast_to(const Value &val, AttrType type, Value &result) const override; + RC cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion = true) const override; RC set_value_from_str(Value &val, const string &data) const override; int cast_cost(AttrType type) override; RC to_string(const Value &val, string &result) const override; -}; \ No newline at end of file +}; diff --git a/src/observer/common/type/data_type.cpp b/src/observer/common/type/data_type.cpp index 5dc3d410..561b8269 100644 --- a/src/observer/common/type/data_type.cpp +++ b/src/observer/common/type/data_type.cpp @@ -12,6 +12,8 @@ See the Mulan PSL v2 for more details. */ #include "common/type/float_type.h" #include "common/type/integer_type.h" #include "common/type/data_type.h" +#include "common/type/boolean_type.h" +#include "common/type/null_type.h" #include "common/type/vector_type.h" array, static_cast(AttrType::MAXTYPE)> DataType::type_instances_ = { @@ -19,6 +21,9 @@ array, static_cast(AttrType::MAXTYPE)> DataType::type_ make_unique(), make_unique(), make_unique(), + make_unique(), + make_unique(), + make_unique(), + make_unique(), make_unique(), - make_unique(AttrType::BOOLEANS), -}; \ No newline at end of file +}; diff --git a/src/observer/common/type/data_type.h b/src/observer/common/type/data_type.h index 7067df4e..2e4508ec 100644 --- a/src/observer/common/type/data_type.h +++ b/src/observer/common/type/data_type.h @@ -75,7 +75,10 @@ class DataType /** * @brief 将 val 转换为 type 类型，并将结果保存到 result 中 */ - virtual RC cast_to(const Value &val, AttrType type, Value &result) const { return RC::UNSUPPORTED; } + virtual RC cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion = true) const + { + return RC::UNSUPPORTED; + } /** * @brief 将 val 转换为 string，并将结果保存到 result 中 diff --git a/src/observer/common/type/date_type.cpp b/src/observer/common/type/date_type.cpp new file mode 100644 index 00000000..3ff5f1e6 --- /dev/null +++ b/src/observer/common/type/date_type.cpp @@ -0,0 +1,49 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/11 * + * @Description : DateType source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "date_type.h" + +#include +#include +#include +#include + +int DateType::compare(const Value &left, const Value &right) const +{ + return common::compare_int((void *)&left.value_.int_value_, (void *)&right.value_.int_value_); +} + +RC DateType::to_string(const Value &val, string &result) const +{ + // 提取年、月、日 + int year = val.value_.int_value_ / 10000; // 获取年份 + int month = (val.value_.int_value_ / 100) % 100; // 获取月份 + int day = val.value_.int_value_ % 100; // 获取日期 + + // 使用字符串流构建结果字符串 + std::ostringstream oss; + oss << std::setw(4) << std::setfill('0') << year << '-' << std::setw(2) << std::setfill('0') << month << '-' + << std::setw(2) << std::setfill('0') << day; + + // 将结果赋值给 result + result = oss.str(); + return RC::SUCCESS; +} + +int DateType::cast_cost(AttrType type) +{ + if (type == AttrType::DATES) + return 0; // DATE -> DATE + if (type == AttrType::INTS) + return 0; // DATE -> INT (需转换) + return INT32_MAX; // 不支持转换 +} diff --git a/src/observer/common/type/date_type.h b/src/observer/common/type/date_type.h new file mode 100644 index 00000000..841179c0 --- /dev/null +++ b/src/observer/common/type/date_type.h @@ -0,0 +1,32 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/11 * + * @Description : DateType header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "common/type/data_type.h" + +/** + * @brief 日期类型 + * @ingroup DateType + */ +class DateType : public DataType +{ +public: + DateType() : DataType(AttrType::DATES) {} + ~DateType() override = default; + + // 是否需要考虑日期与其它类型的转换？（不需要） + // 不用实现 cast to，也不需要考虑其他类型转 date + int cast_cost(AttrType type) override; + int compare(const Value &left, const Value &right) const override; + RC to_string(const Value &val, string &result) const override; +}; diff --git a/src/observer/common/type/float_type.cpp b/src/observer/common/type/float_type.cpp index fcf89a7c..9c09b5c9 100644 --- a/src/observer/common/type/float_type.cpp +++ b/src/observer/common/type/float_type.cpp @@ -8,6 +8,8 @@ EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. See the Mulan PSL v2 for more details. */ +#include + #include "common/lang/comparator.h" #include "common/lang/sstream.h" #include "common/log/log.h" @@ -44,12 +46,11 @@ RC FloatType::multiply(const Value &left, const Value &right, Value &result) con RC FloatType::divide(const Value &left, const Value &right, Value &result) const { if (right.get_float() > -EPSILON && right.get_float() < EPSILON) { - // NOTE: - // 设置为浮点数最大值是不正确的。通常的做法是设置为NULL，但是当前的miniob没有NULL概念，所以这里设置为浮点数最大值。 - result.set_float(numeric_limits::max()); - } else { - result.set_float(left.get_float() / right.get_float()); + result.set_null(); + return RC::SUCCESS; } + + result.set_float(left.get_float() / right.get_float()); return RC::SUCCESS; } @@ -61,7 +62,7 @@ RC FloatType::negative(const Value &val, Value &result) const RC FloatType::set_value_from_str(Value &val, const string &data) const { - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; stringstream deserialize_stream; deserialize_stream.clear(); deserialize_stream.str(data); @@ -83,3 +84,46 @@ RC FloatType::to_string(const Value &val, string &result) const result = ss.str(); return RC::SUCCESS; } + +int FloatType::cast_cost(AttrType type) +{ + if (type == AttrType::FLOATS) + return 0; // FLOAT -> FLOAT + // if (type == AttrType::INTS) + // return 1; // FLOAT -> INT (可能丢失精度，也不支持转换) + if (type == AttrType::BOOLEANS) + return 1; // FLOAT -> BOOL (非严格转换) + return INT32_MAX; // 不支持转换 +} + +RC FloatType::cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion) const +{ + switch (type) { + case AttrType::FLOATS: { + result.set_float(val.get_float()); + } break; + case AttrType::INTS: { + // 浮点数转整数（四舍五入，如1.3转换为1，1.5转换为2） + // 涉及浮点数的比较，整数转换为浮点数，字符串转换为浮点数，再比较 + // 使用 std::round 进行四舍五入 + result.set_int(static_cast(std::round(val.get_float()))); + } break; + case AttrType::CHARS: { + // 数字转字符串，不带符号，不考虑溢出 + // 浮点数的表示方法是什么?（举例：1.5转换为'1.5'） + float float_val = val.get_float(); + std::string str = float_val < 0 ? std::to_string(-float_val) : std::to_string(float_val); + // 去除尾随的0 和可能的 '.' + str.erase(str.find_last_not_of('0') + 1, std::string::npos); + if (str.back() == '.') { + str.pop_back(); + } + result.set_string(str.c_str()); // 设置字符串结果 + } break; + case AttrType::BOOLEANS: { + result.set_boolean(val.get_float() != 0.0f); // 非零为 true，零为 false + } break; + default: return RC::UNSUPPORTED; // 不支持的转换 + } + return RC::SUCCESS; +} diff --git a/src/observer/common/type/float_type.h b/src/observer/common/type/float_type.h index fd4a7fcb..0064e5c0 100644 --- a/src/observer/common/type/float_type.h +++ b/src/observer/common/type/float_type.h @@ -11,6 +11,7 @@ See the Mulan PSL v2 for more details. */ #pragma once #include "common/type/data_type.h" +#include /** * @brief 浮点型数据类型 @@ -30,7 +31,10 @@ class FloatType : public DataType RC divide(const Value &left, const Value &right, Value &result) const override; RC negative(const Value &val, Value &result) const override; + int cast_cost(AttrType type) override; + RC cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion = true) const override; + RC set_value_from_str(Value &val, const string &data) const override; RC to_string(const Value &val, string &result) const override; -}; \ No newline at end of file +}; diff --git a/src/observer/common/type/integer_type.cpp b/src/observer/common/type/integer_type.cpp index 8188cd0b..65fd6306 100644 --- a/src/observer/common/type/integer_type.cpp +++ b/src/observer/common/type/integer_type.cpp @@ -20,7 +20,8 @@ int IntegerType::compare(const Value &left, const Value &right) const ASSERT(right.attr_type() == AttrType::INTS || right.attr_type() == AttrType::FLOATS, "right type is not numeric"); if (right.attr_type() == AttrType::INTS) { return common::compare_int((void *)&left.value_.int_value_, (void *)&right.value_.int_value_); - } else if (right.attr_type() == AttrType::FLOATS) { + } + if (right.attr_type() == AttrType::FLOATS) { float left_val = left.get_float(); float right_val = right.get_float(); return common::compare_float((void *)&left_val, (void *)&right_val); @@ -54,7 +55,7 @@ RC IntegerType::negative(const Value &val, Value &result) const RC IntegerType::set_value_from_str(Value &val, const string &data) const { - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; stringstream deserialize_stream; deserialize_stream.clear(); // 清理stream的状态，防止多次解析出现异常 deserialize_stream.str(data); @@ -70,8 +71,43 @@ RC IntegerType::set_value_from_str(Value &val, const string &data) const RC IntegerType::to_string(const Value &val, string &result) const { - stringstream ss; - ss << val.value_.int_value_; - result = ss.str(); + result = std::to_string(val.get_int()); + return RC::SUCCESS; +} + +int IntegerType::cast_cost(AttrType type) +{ + if (type == AttrType::INTS) + return 0; // INT -> INT + if (type == AttrType::FLOATS) + return 1; // INT -> FLOAT + if (type == AttrType::BOOLEANS) + return 1; // INT -> BOOL (非严格转换) + return INT32_MAX; // 不支持转换 +} + +RC IntegerType::cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion) const +{ + switch (type) { + case AttrType::INTS: { + result.set_int(val.get_int()); + } break; + case AttrType::FLOATS: { + // 整数转浮点数，直接转换为浮点数，不考虑溢出 + result.set_float(static_cast(val.get_int())); + } break; + case AttrType::BOOLEANS: { + // 0 为 false，其他为 true + result.set_boolean(val.get_int() != 0); + } break; + case AttrType::CHARS: { + // 数字转字符串，不带符号，不考虑长度溢出 + int int_val = val.get_int(); + std::string str = int_val < 0 ? std::to_string(-int_val) : std::to_string(int_val); + result.set_string(str.c_str()); // 设置字符串结果 + break; + } break; + default: return RC::UNSUPPORTED; // 不支持的转换 + } return RC::SUCCESS; -} \ No newline at end of file +} diff --git a/src/observer/common/type/integer_type.h b/src/observer/common/type/integer_type.h index dd8c48e5..75dcd6de 100644 --- a/src/observer/common/type/integer_type.h +++ b/src/observer/common/type/integer_type.h @@ -12,6 +12,8 @@ See the Mulan PSL v2 for more details. */ #include "common/type/data_type.h" +#include + /** * @brief 整型类型 * @ingroup DataType @@ -32,4 +34,7 @@ class IntegerType : public DataType RC set_value_from_str(Value &val, const string &data) const override; RC to_string(const Value &val, string &result) const override; -}; \ No newline at end of file + + int cast_cost(AttrType type) override; + RC cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion = true) const override; +}; diff --git a/src/observer/sql/expr/aggregator.cpp b/src/observer/common/type/null_type.cpp similarity index 51% rename from src/observer/sql/expr/aggregator.cpp rename to src/observer/common/type/null_type.cpp index 29ba468a..981fdb17 100644 --- a/src/observer/sql/expr/aggregator.cpp +++ b/src/observer/common/type/null_type.cpp @@ -8,29 +8,28 @@ EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. See the Mulan PSL v2 for more details. */ -// -// Created by Wangyunlai on 2024/05/29. -// - -#include "sql/expr/aggregator.h" +#include "common/lang/comparator.h" +#include "common/lang/sstream.h" #include "common/log/log.h" +#include "null_type.h" +#include "common/value.h" + +int NullType::compare(const Value &left, const Value &right) const +{ + ASSERT(left.attr_type() == AttrType::NULLS, "left type is not a null"); + return -1; +} -RC SumAggregator::accumulate(const Value &value) +RC NullType::to_string(const Value &val, string &result) const { - if (value_.attr_type() == AttrType::UNDEFINED) { - value_ = value; - return RC::SUCCESS; - } - - ASSERT(value.attr_type() == value_.attr_type(), "type mismatch. value type: %s, value_.type: %s", - attr_type_to_string(value.attr_type()), attr_type_to_string(value_.attr_type())); - - Value::add(value, value_, value_); + result = "NULL"; return RC::SUCCESS; } -RC SumAggregator::evaluate(Value& result) +RC NullType::cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion) const { - result = value_; + ASSERT(val.attr_type() == AttrType::NULLS, "val type is not a null"); + result.set_type(type); + result.set_null(); return RC::SUCCESS; } diff --git a/src/observer/common/type/null_type.h b/src/observer/common/type/null_type.h new file mode 100644 index 00000000..c90bc02e --- /dev/null +++ b/src/observer/common/type/null_type.h @@ -0,0 +1,27 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +#pragma once + +#include "common/type/data_type.h" + +class NullType : public DataType +{ +public: + NullType() : DataType(AttrType::NULLS) {} + + ~NullType() override {} + + int compare(const Value &left, const Value &right) const override; + + RC to_string(const Value &val, string &result) const override; + + RC cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion = true) const override; +}; diff --git a/src/observer/common/type/text_type.cpp b/src/observer/common/type/text_type.cpp new file mode 100644 index 00000000..d78eaacd --- /dev/null +++ b/src/observer/common/type/text_type.cpp @@ -0,0 +1,106 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/30 * + * @Description : TextType source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "common/utils.h" +#include "common/type/text_type.h" +#include "common/value.h" +#include "common/log/log.h" +#include "common/lang/comparator.h" + +int TextType::compare(const Value &left, const Value &right) const +{ + ASSERT(left.attr_type() == AttrType::TEXTS && right.attr_type() == AttrType::TEXTS, "invalid type"); + return common::compare_string( + (void *)left.value_.pointer_value_, left.length_, (void *)right.value_.pointer_value_, right.length_); +} + +int TextType::cast_cost(AttrType type) +{ + if (type == AttrType::CHARS || type == AttrType::TEXTS) { + return 0; + } + if (type == AttrType::DATES) { + return 1; + } + if (type == AttrType::INTS) { + return 1; + } + if (type == AttrType::FLOATS) { + return 1; + } + return INT32_MAX; +} + +RC TextType::cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion) const +{ + switch (type) { + case AttrType::TEXTS: { + result.set_text(val.value_.pointer_value_, val.length_); + } break; + case AttrType::DATES: { + int date_val; + RC rc = parse_date(val.value_.pointer_value_, date_val); + if (rc != RC::SUCCESS) { + return rc; + } + result.set_date(date_val); + } break; + // 字符串转数字 + // 如果字符串刚好是一个数字，则转换为对应的数字（如'1'转换为1，'2.1'转换为2.1） + // 如果字符串的前缀是一个数字，则转换前缀数字部分为数字（如'1a1'转换为1，'2.1a'转换为2.1） + // 如果字符串前缀不是任何合法的数字，则转换为0（不需要考虑前导符号 '+' '-'） + // 如果转换数字溢出怎么处理?（不考虑） + // 是否考虑十六进制/八进制?（不考虑） + case AttrType::INTS: { + // WHERE id < '1.5a'; + // 先当浮点数解析，浮点数可以兼容整型，再四舍五入转 int + float float_val; + RC rc = parse_float_prefix(val.value_.pointer_value_, float_val); + if (rc != RC::SUCCESS) { + return rc; + } + int int_val = static_cast(float_val); + // 是整数 + if (int_val == float_val || !allow_type_promotion) { + result.set_int(int_val); + } else { + // 为浮点数，类型提升 + result.set_float(float_val); + } + } break; + case AttrType::FLOATS: { + float float_val; + RC rc = parse_float_prefix(val.value_.pointer_value_, float_val); + if (rc != RC::SUCCESS) { + return rc; + } + result.set_float(float_val); + } break; + default: return RC::UNIMPLEMENTED; + } + return RC::SUCCESS; +} + +RC TextType::to_string(const Value &val, string &result) const +{ + stringstream ss; + ss << val.value_.pointer_value_; + result = ss.str(); + return RC::SUCCESS; +} + +// 该函数会被 load 使用 +RC TextType::set_value_from_str(Value &val, const string &data) const +{ + val.set_string(data.c_str()); + return RC::SUCCESS; +} diff --git a/src/observer/common/type/text_type.h b/src/observer/common/type/text_type.h new file mode 100644 index 00000000..69ed7715 --- /dev/null +++ b/src/observer/common/type/text_type.h @@ -0,0 +1,38 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/30 * + * @Description : TextType header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "common/rc.h" +#include "common/type/data_type.h" + +/** + * @brief 固定长度的字符串类型 + * @ingroup DataType + */ +class TextType : public DataType +{ +public: + TextType() : DataType(AttrType::TEXTS) {} + + ~TextType() override = default; + + int compare(const Value &left, const Value &right) const override; + + RC cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion = true) const override; + + RC set_value_from_str(Value &val, const string &data) const override; + + int cast_cost(AttrType type) override; + + RC to_string(const Value &val, string &result) const override; +}; diff --git a/src/observer/common/type/vector_type.cpp b/src/observer/common/type/vector_type.cpp new file mode 100644 index 00000000..8389ac28 --- /dev/null +++ b/src/observer/common/type/vector_type.cpp @@ -0,0 +1,161 @@ +// +// Created by root on 24-10-14. +// + +#include "common/type/vector_type.h" + +#include "common/value.h" +#include + +int VectorType::compare(const Value &left, const Value &right) const +{ + // 获取左向量和右向量的指针 + const float *left_data = reinterpret_cast(left.data()); + const float *right_data = reinterpret_cast(right.data()); + + // 获取向量的长度 + int left_count = left.length() / sizeof(float); + int right_count = right.length() / sizeof(float); + + // 如果向量长度不一致，先比较长度 + if (left_count != right_count) { + return (left_count < right_count) ? -1 : 1; + } + + // 逐元素比较两个向量 + for (int i = 0; i < left_count; ++i) { + if (left_data[i] < right_data[i]) { + return -1; // left 小于 right + } else if (left_data[i] > right_data[i]) { + return 1; // left 大于 right + } + } + + // 如果每个元素都相等，则两个向量相等 + return 0; +} + +RC VectorType::cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion) const +{ + if (type == AttrType::CHARS) { + result = Value(val.to_string().c_str()); + return RC::SUCCESS; + } + return RC::INTERNAL; +} +RC VectorType::set_value_from_str(Value &val, const string &data) const { return RC::UNIMPLEMENTED; } +int VectorType::cast_cost(AttrType type) { return DataType::cast_cost(type); } +RC VectorType::to_string(const Value &val, std::string &result) const +{ + // 获取 float 数组指针 + const auto *data = reinterpret_cast(val.data()); + + // 计算数组中元素的数量 + int count = val.length() / sizeof(float); + + // 使用字符串流来构建输出字符串 + std::ostringstream oss; + oss << "["; + + // 遍历数组元素并拼接成字符串 + for (int i = 0; i < count; ++i) { + if (i != 0) { + oss << ","; // 在每个元素之间加逗号和空格 + } + oss << common::double_to_str(data[i]); + } + + oss << "]"; // 关闭数组的方括号 + + // 将结果存储到 result 中 + result = oss.str(); + + return RC::SUCCESS; +} + +RC VectorType::add(const Value &left, const Value &right, Value &result) const +{ + // 获取左向量和右向量的指针 + const float *left_data = reinterpret_cast(left.data()); + const float *right_data = reinterpret_cast(right.data()); + + // 获取向量的长度 + int left_count = left.length() / sizeof(float); + int right_count = right.length() / sizeof(float); + + // 确保两个向量的长度相同 + if (left_count != right_count) { + return RC::VECTOR_DIM_MISMATCH; // 长度不匹配，返回错误 + } + + // 创建一个新的结果向量 + float *result_data = new float[left_count]; + + // 执行逐元素加法 + for (int i = 0; i < left_count; ++i) { + result_data[i] = left_data[i] + right_data[i]; + } + + // 将结果存储到 result 中 + result.set_vector(result_data, left.length()); + + return RC::SUCCESS; +} + +RC VectorType::subtract(const Value &left, const Value &right, Value &result) const +{ + // 获取左向量和右向量的指针 + const float *left_data = reinterpret_cast(left.data()); + const float *right_data = reinterpret_cast(right.data()); + + // 获取向量的长度 + int left_count = left.length() / sizeof(float); + int right_count = right.length() / sizeof(float); + + // 确保两个向量的长度相同 + if (left_count != right_count) { + return RC::VECTOR_DIM_MISMATCH; // 长度不匹配，返回错误 + } + + // 创建一个新的结果向量 + float *result_data = new float[left_count]; + + // 执行逐元素减法 + for (int i = 0; i < left_count; ++i) { + result_data[i] = left_data[i] - right_data[i]; + } + + // 将结果存储到 result 中 + result.set_vector(result_data, left.length()); + + return RC::SUCCESS; +} + +RC VectorType::multiply(const Value &left, const Value &right, Value &result) const +{ + // 获取左向量和右向量的指针 + const float *left_data = reinterpret_cast(left.data()); + const float *right_data = reinterpret_cast(right.data()); + + // 获取向量的长度 + int left_count = left.length() / sizeof(float); + int right_count = right.length() / sizeof(float); + + // 确保两个向量的长度相同 + if (left_count != right_count) { + return RC::VECTOR_DIM_MISMATCH; // 长度不匹配，返回错误 + } + + // 创建一个新的结果向量 + float *result_data = new float[left_count]; + + // 执行逐元素乘法 + for (int i = 0; i < left_count; ++i) { + result_data[i] = left_data[i] * right_data[i]; + } + + // 将结果存储到 result 中 + result.set_vector(result_data, left.length()); + + return RC::SUCCESS; +} diff --git a/src/observer/common/type/vector_type.h b/src/observer/common/type/vector_type.h index 9e9f8150..8ea6578e 100644 --- a/src/observer/common/type/vector_type.h +++ b/src/observer/common/type/vector_type.h @@ -20,13 +20,22 @@ class VectorType : public DataType { public: VectorType() : DataType(AttrType::VECTORS) {} - virtual ~VectorType() {} - int compare(const Value &left, const Value &right) const override { return INT32_MAX; } + virtual ~VectorType() = default; - RC add(const Value &left, const Value &right, Value &result) const override { return RC::UNIMPLEMENTED; } - RC subtract(const Value &left, const Value &right, Value &result) const override { return RC::UNIMPLEMENTED; } - RC multiply(const Value &left, const Value &right, Value &result) const override { return RC::UNIMPLEMENTED; } + int compare(const Value &left, const Value &right) const override; - RC to_string(const Value &val, string &result) const override { return RC::UNIMPLEMENTED; } -}; \ No newline at end of file + RC cast_to(const Value &val, AttrType type, Value &result, bool allow_type_promotion = true) const override; + + RC set_value_from_str(Value &val, const string &data) const override; + + int cast_cost(AttrType type) override; + + RC to_string(const Value &val, string &result) const override; + + RC add(const Value &left, const Value &right, Value &result) const override; + + RC subtract(const Value &left, const Value &right, Value &result) const override; + + RC multiply(const Value &left, const Value &right, Value &result) const override; +}; diff --git a/src/observer/common/utils.cpp b/src/observer/common/utils.cpp new file mode 100644 index 00000000..acd0be05 --- /dev/null +++ b/src/observer/common/utils.cpp @@ -0,0 +1,102 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/30 * + * @Description : utils source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "utils.h" + +#include +#include +#include +#include + +bool check_date(int y, int m, int d) +{ + // 定义每个月的天数 + static int mon[] = {0, 31, 28, 31, 30, 31, 30, 31, 31, 30, 31, 30, 31}; + // 判断是否为闰年 + bool leap = (y % 400 == 0) || (y % 100 != 0 && y % 4 == 0); + // 检查年份、月份和日期的合法性 + return (y > 0 && y <= 9999) && (m > 0 && m <= 12) && + (d > 0 && d <= (mon[m] + (m == 2 && leap ? 1 : 0))); // 2月如果是闰年则加1天 +} + +RC parse_date(const char *str, int &result) +{ + int y, m, d; + if (sscanf(str, "%d-%d-%d", &y, &m, &d) != 3) { + return RC::INVALID_ARGUMENT; + } + if (!check_date(y, m, d)) { + return RC::INVALID_ARGUMENT; + } + result = y * 10000 + m * 100 + d; + return RC::SUCCESS; +} + +RC parse_float_prefix(const char *str, float &result) +{ + char *end_ptr; + // 输入不是有效的整数格式，会转换为 0.0 + double float_val = std::strtod(str, &end_ptr); + // 注释以支持前缀解析 + // if (*end_ptr != '\0' && !isspace(*end_ptr)) { + // return RC::INVALID_ARGUMENT; // 浮点数后应为结尾或空白字符 + // } + // 默认认为 float_val 是否在 float 范围内 + result = static_cast(float_val); + return RC::SUCCESS; +} + +RC parse_vector_from_string(const char *str, float *&array, int &length) +{ + if (!str || *str != '[') { + return RC::INVALID_ARGUMENT; + } + + std::string s = str; + if (s.back() != ']') { + return RC::INVALID_ARGUMENT; + } + + s = s.substr(1, s.size() - 2); // 去掉开头和结尾的方括号 + std::stringstream ss(s); + std::string token; + size_t count = 0; + + // 先统计有多少个浮点数 + while (std::getline(ss, token, ',')) { + count++; + } + + if (count == 0) { + return RC::INVALID_ARGUMENT; // 空数组 + } + + // 分配数组内存 + array = new float[count]; + length = count * sizeof(float); + + // 重置流并解析浮点数 + ss.clear(); + ss.str(s); + size_t index = 0; + + while (std::getline(ss, token, ',')) { + std::stringstream valueStream(token); + if (!(valueStream >> array[index])) { + delete[] array; // 清理已分配内存 + return RC::VECTOR_PARSE_ERROR; + } + index++; + } + + return RC::SUCCESS; +} diff --git a/src/observer/common/utils.h b/src/observer/common/utils.h new file mode 100644 index 00000000..a184b255 --- /dev/null +++ b/src/observer/common/utils.h @@ -0,0 +1,24 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/30 * + * @Description : utils header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "rc.h" +#include + +bool check_date(int y, int m, int d); + +RC parse_date(const char *str, int &result); + +RC parse_float_prefix(const char *str, float &result); + +RC parse_vector_from_string(const char *str, float *&array, int &length); diff --git a/src/observer/common/value.cpp b/src/observer/common/value.cpp index b8f04914..c61bd070 100644 --- a/src/observer/common/value.cpp +++ b/src/observer/common/value.cpp @@ -12,6 +12,8 @@ See the Mulan PSL v2 for more details. */ // Created by WangYunlai on 2023/06/28. // +#include + #include "common/value.h" #include "common/lang/comparator.h" @@ -20,27 +22,42 @@ See the Mulan PSL v2 for more details. */ #include "common/lang/string.h" #include "common/log/log.h" +Value::Value(NullValue) : attr_type_(AttrType::NULLS) { set_null(); } + Value::Value(int val) { set_int(val); } Value::Value(float val) { set_float(val); } Value::Value(bool val) { set_boolean(val); } -Value::Value(const char *s, int len /*= 0*/) { set_string(s, len); } +Value::Value(const char *s, int len /*= 0*/) +{ + if (OB_SUCC(parse_vector_from_string(s, value_.vector_value_, length_))) { + set_vector(); + } else { + set_string(s, len); + } +} + +Value::Value(const vector &values) { set_vector(values); } Value::Value(const Value &other) { this->attr_type_ = other.attr_type_; this->length_ = other.length_; this->own_data_ = other.own_data_; - switch (this->attr_type_) { - case AttrType::CHARS: { - set_string_from_other(other); - } break; - - default: { - this->value_ = other.value_; - } break; + this->is_null_ = other.is_null_; + // 如果是 null 值没必要拷贝 value，因为此时 value 是垃圾值 + if (!this->is_null_) { + switch (this->attr_type_) { + case AttrType::CHARS: + case AttrType::TEXTS: { + set_string_from_other(other); + } break; + default: { + this->value_ = other.value_; + } break; + } } } @@ -49,9 +66,12 @@ Value::Value(Value &&other) this->attr_type_ = other.attr_type_; this->length_ = other.length_; this->own_data_ = other.own_data_; - this->value_ = other.value_; - other.own_data_ = false; - other.length_ = 0; + this->is_null_ = other.is_null_; + if (!this->is_null_) { + this->value_ = other.value_; + } + other.own_data_ = false; + other.length_ = 0; } Value &Value::operator=(const Value &other) @@ -63,14 +83,17 @@ Value &Value::operator=(const Value &other) this->attr_type_ = other.attr_type_; this->length_ = other.length_; this->own_data_ = other.own_data_; - switch (this->attr_type_) { - case AttrType::CHARS: { - set_string_from_other(other); - } break; - - default: { - this->value_ = other.value_; - } break; + this->is_null_ = other.is_null_; + if (!this->is_null_) { + switch (this->attr_type_) { + case AttrType::CHARS: + case AttrType::TEXTS: { + set_string_from_other(other); + } break; + default: { + this->value_ = other.value_; + } break; + } } return *this; } @@ -84,9 +107,12 @@ Value &Value::operator=(Value &&other) this->attr_type_ = other.attr_type_; this->length_ = other.length_; this->own_data_ = other.own_data_; - this->value_ = other.value_; - other.own_data_ = false; - other.length_ = 0; + this->is_null_ = other.is_null_; + if (!this->is_null_) { + this->value_ = other.value_; + } + other.own_data_ = false; + other.length_ = 0; return *this; } @@ -94,17 +120,19 @@ void Value::reset() { switch (attr_type_) { case AttrType::CHARS: + case AttrType::TEXTS: { if (own_data_ && value_.pointer_value_ != nullptr) { delete[] value_.pointer_value_; value_.pointer_value_ = nullptr; } - break; + } break; default: break; } attr_type_ = AttrType::UNDEFINED; length_ = 0; own_data_ = false; + is_null_ = false; } void Value::set_data(char *data, int length) @@ -113,6 +141,9 @@ void Value::set_data(char *data, int length) case AttrType::CHARS: { set_string(data, length); } break; + case AttrType::TEXTS: { + set_text(data, length); + } break; case AttrType::INTS: { value_.int_value_ = *(int *)data; length_ = length; @@ -125,6 +156,14 @@ void Value::set_data(char *data, int length) value_.bool_value_ = *(int *)data != 0; length_ = length; } break; + case AttrType::DATES: { + value_.int_value_ = *(int *)data; + length_ = length; + } break; + case AttrType::VECTORS: { + value_.vector_value_ = (float *)data; + length_ = length; + } break; default: { LOG_WARN("unknown data type: %d", attr_type_); } break; @@ -146,6 +185,7 @@ void Value::set_float(float val) value_.float_value_ = val; length_ = sizeof(val); } + void Value::set_boolean(bool val) { reset(); @@ -154,6 +194,18 @@ void Value::set_boolean(bool val) length_ = sizeof(val); } +void Value::set_date(int val) +{ + reset(); + attr_type_ = AttrType::DATES; + value_.int_value_ = val; + length_ = sizeof(val); +} +void Value::set_vector() +{ + attr_type_ = AttrType::VECTORS; + own_data_ = true; +} void Value::set_string(const char *s, int len /*= 0*/) { reset(); @@ -175,6 +227,48 @@ void Value::set_string(const char *s, int len /*= 0*/) } } +void Value::set_text(const char *s, int len /*= 65535*/) +{ + reset(); + attr_type_ = AttrType::TEXTS; + if (s == nullptr) { + value_.pointer_value_ = nullptr; + length_ = 0; + } else { + own_data_ = true; + if (len > 0) { + len = strnlen(s, len); + } else { + len = strlen(s); + } + value_.pointer_value_ = new char[len + 1]; + length_ = len; + memcpy(value_.pointer_value_, s, len); + value_.pointer_value_[len] = '\0'; + } +} + +void Value::set_vector(float *array, int length) +{ + attr_type_ = AttrType::VECTORS; + length_ = length; + value_.vector_value_ = array; + + own_data_ = true; +} + +void Value::set_vector(const vector &val) +{ + attr_type_ = AttrType::VECTORS; + length_ = val.size() * sizeof(float); + value_.vector_value_ = new float[length_]; + for (size_t i = 0; i < val.size(); i++) { + value_.vector_value_[i] = val[i]; + } + + own_data_ = true; +} + void Value::set_value(const Value &value) { switch (value.attr_type_) { @@ -190,6 +284,12 @@ void Value::set_value(const Value &value) case AttrType::BOOLEANS: { set_boolean(value.get_boolean()); } break; + case AttrType::DATES: { + set_date(value.get_int()); + } break; + case AttrType::TEXTS: { + set_text(value.get_string().c_str()); + } break; default: { ASSERT(false, "got an invalid value type"); } break; @@ -198,8 +298,8 @@ void Value::set_value(const Value &value) void Value::set_string_from_other(const Value &other) { - ASSERT(attr_type_ == AttrType::CHARS, "attr type is not CHARS"); - if (own_data_ && other.value_.pointer_value_ != nullptr && length_ != 0) { + ASSERT(attr_type_ == AttrType::CHARS || attr_type() == AttrType::TEXTS, "attr type is not CHARS or TEXTS"); + if (own_data_ && other.value_.pointer_value_ != nullptr) { this->value_.pointer_value_ = new char[this->length_ + 1]; memcpy(this->value_.pointer_value_, other.value_.pointer_value_, this->length_); this->value_.pointer_value_[this->length_] = '\0'; @@ -209,9 +309,13 @@ void Value::set_string_from_other(const Value &other) const char *Value::data() const { switch (attr_type_) { - case AttrType::CHARS: { + case AttrType::CHARS: + case AttrType::TEXTS: { return value_.pointer_value_; } break; + case AttrType::VECTORS: { + return (char *)value_.vector_value_; + } break; default: { return (const char *)&value_; } break; @@ -220,6 +324,9 @@ const char *Value::data() const string Value::to_string() const { + if (is_null_) { + return "NULL"; + } string res; RC rc = DataType::type_instance(this->attr_type_)->to_string(*this, res); if (OB_FAIL(rc)) { @@ -229,12 +336,29 @@ string Value::to_string() const return res; } -int Value::compare(const Value &other) const { return DataType::type_instance(this->attr_type_)->compare(*this, other); } +int Value::compare(const Value &other) const +{ + return DataType::type_instance(this->attr_type_)->compare(*this, other); +} + +bool Value::LIKE(const Value &other) const +{ + const std::string left_str = this->get_string(); + const std::string &right_str = other.get_string(); + + // 将 SQL 通配符转换为正则表达式 + std::string regex_str = std::regex_replace(right_str, std::regex("%"), ".*"); + regex_str = std::regex_replace(regex_str, std::regex("_"), "."); + + std::regex regex_pattern(regex_str); + return std::regex_match(left_str, regex_pattern); +} int Value::get_int() const { switch (attr_type_) { - case AttrType::CHARS: { + case AttrType::CHARS: + case AttrType::TEXTS: { try { return (int)(std::stol(value_.pointer_value_)); } catch (exception const &ex) { @@ -251,6 +375,11 @@ int Value::get_int() const case AttrType::BOOLEANS: { return (int)(value_.bool_value_); } + case AttrType::DATES: { + // 虽然 date 类型是用 int 存的，但是目前不需要 get_date 函数 + LOG_TRACE("failed to convert date to number. s=%d", value_.int_value_); + return 0; + } default: { LOG_WARN("unknown data type. type=%d", attr_type_); return 0; @@ -262,7 +391,8 @@ int Value::get_int() const float Value::get_float() const { switch (attr_type_) { - case AttrType::CHARS: { + case AttrType::CHARS: + case AttrType::TEXTS: { try { return std::stof(value_.pointer_value_); } catch (exception const &ex) { @@ -279,6 +409,10 @@ float Value::get_float() const case AttrType::BOOLEANS: { return float(value_.bool_value_); } break; + case AttrType::DATES: { + LOG_TRACE("failed to convert date to float. s=%d", value_.int_value_); + return 0; + } default: { LOG_WARN("unknown data type. type=%d", attr_type_); return 0; @@ -292,7 +426,8 @@ string Value::get_string() const { return this->to_string(); } bool Value::get_boolean() const { switch (attr_type_) { - case AttrType::CHARS: { + case AttrType::CHARS: + case AttrType::TEXTS: { try { float val = std::stof(value_.pointer_value_); if (val >= EPSILON || val <= -EPSILON) { @@ -320,6 +455,10 @@ bool Value::get_boolean() const case AttrType::BOOLEANS: { return value_.bool_value_; } break; + case AttrType::DATES: { + LOG_TRACE("failed to convert date to boolean. s=%d", value_.int_value_); + return false; + } default: { LOG_WARN("unknown data type. type=%d", attr_type_); return false; @@ -327,3 +466,47 @@ bool Value::get_boolean() const } return false; } + +int Value::get_date() const +{ + if (attr_type_ == AttrType::DATES) { + return value_.int_value_; + } else { + return 0; + } +} + +RC Value::borrow_text(const Value &v) +{ + ASSERT(v.attr_type() != AttrType::TEXTS, "Only text type can be borrowed! "); + if (v.length_ > 65535) { + return RC::VALUE_TOO_LONG; + } + reset(); + attr_type_ = AttrType::TEXTS; + value_.pointer_value_ = v.value_.pointer_value_; + length_ = v.length_; + return RC::SUCCESS; +} + +std::vector Value::get_vector() const +{ + assert(attr_type_ == AttrType::VECTORS); + std::vector vector(get_vector_length()); + for (int i = 0; i < vector.size(); ++i) { + vector[i] = get_vector_element(i); + } + return vector; +} + +int Value::get_vector_length() const +{ + assert(attr_type_ == AttrType::VECTORS); + return length_ / sizeof(float); +} + +float Value::get_vector_element(int i) const +{ + assert(attr_type_ == AttrType::VECTORS); + return value_.vector_value_[i]; +} diff --git a/src/observer/common/value.h b/src/observer/common/value.h index 3c0481a6..8b43f10f 100644 --- a/src/observer/common/value.h +++ b/src/observer/common/value.h @@ -18,6 +18,13 @@ See the Mulan PSL v2 for more details. */ #include "common/lang/memory.h" #include "common/type/attr_type.h" #include "common/type/data_type.h" +#include "common/type/date_type.h" +#include "common/type/text_type.h" +#include "common/type/boolean_type.h" +#include "common/utils.h" + +class NullValue +{}; /** * @brief 属性的值 @@ -34,6 +41,9 @@ class Value final friend class FloatType; friend class BooleanType; friend class CharType; + friend class DateType; + friend class NullType; + friend class TextType; friend class VectorType; Value() = default; @@ -42,10 +52,12 @@ class Value final Value(AttrType attr_type, char *data, int length = 4) : attr_type_(attr_type) { this->set_data(data, length); } + explicit Value(NullValue); explicit Value(int val); explicit Value(float val); explicit Value(bool val); explicit Value(const char *s, int len = 0); + explicit Value(const vector &values); Value(const Value &other); Value(Value &&other); @@ -80,9 +92,9 @@ class Value final return DataType::type_instance(result.attr_type())->negative(value, result); } - static RC cast_to(const Value &value, AttrType to_type, Value &result) + static RC cast_to(const Value &value, AttrType to_type, Value &result, bool allow_type_promotion = true) { - return DataType::type_instance(value.attr_type())->cast_to(value, to_type, result); + return DataType::type_instance(value.attr_type())->cast_to(value, to_type, result, allow_type_promotion); } void set_type(AttrType type) { this->attr_type_ = type; } @@ -90,30 +102,53 @@ class Value final void set_data(const char *data, int length) { this->set_data(const_cast(data), length); } void set_value(const Value &value); void set_boolean(bool val); + void set_null(bool is_null = true) { is_null_ = is_null; } string to_string() const; - int compare(const Value &other) const; + int compare(const Value &other) const; + bool LIKE(const Value &other) const; const char *data() const; int length() const { return length_; } AttrType attr_type() const { return attr_type_; } + RC borrow_text(const Value &v); + public: /** * 获取对应的值 * 如果当前的类型与期望获取的类型不符，就会执行转换操作 */ - int get_int() const; - float get_float() const; - string get_string() const; - bool get_boolean() const; + int get_int() const; + float get_float() const; + string get_string() const; + bool get_boolean() const; + int get_date() const; + std::vector get_vector() const; + int get_vector_length() const; + float get_vector_element(int i) const; + bool is_null() const { return is_null_; } + inline bool is_str() const { return attr_type_ == AttrType::CHARS; } + + static int implicit_cast_cost(AttrType from, AttrType to) + { + if (from == to) { + return 0; + } + return DataType::type_instance(from)->cast_cost(to); + } private: void set_int(int val); void set_float(float val); + void set_date(int val); void set_string(const char *s, int len = 0); + void set_text(const char *s, int len = 65535); + void set_vector(float *array, int length); + void set_vector(const vector &val); + void set_vector(); void set_string_from_other(const Value &other); private: @@ -126,8 +161,12 @@ class Value final float float_value_; bool bool_value_; char *pointer_value_; + float *vector_value_; } value_ = {.int_value_ = 0}; + static_assert(sizeof(std::vector *) == sizeof(char *)); + /// 是否申请并占有内存, 目前对于 CHARS 类型 own_data_ 为true, 其余类型 own_data_ 为false bool own_data_ = false; + bool is_null_ = false; }; diff --git a/src/observer/event/sql_debug.cpp b/src/observer/event/sql_debug.cpp index ab68dc17..185d4768 100644 --- a/src/observer/event/sql_debug.cpp +++ b/src/observer/event/sql_debug.cpp @@ -18,7 +18,6 @@ See the Mulan PSL v2 for more details. */ #include "event/sql_debug.h" #include "session/session.h" - void SqlDebug::add_debug_info(const string &debug_info) { debug_infos_.push_back(debug_info); } void SqlDebug::clear_debug_info() { debug_infos_.clear(); } diff --git a/src/observer/net/cli_communicator.cpp b/src/observer/net/cli_communicator.cpp index 396dc280..432cda89 100644 --- a/src/observer/net/cli_communicator.cpp +++ b/src/observer/net/cli_communicator.cpp @@ -28,7 +28,7 @@ See the Mulan PSL v2 for more details. */ #include "readline/readline.h" #endif -#define MAX_MEM_BUFFER_SIZE 8192 +#define MAX_MEM_BUFFER_SIZE 131072 #define PORT_DEFAULT 6789 using namespace common; @@ -39,7 +39,8 @@ time_t last_history_write_time = 0; sigjmp_buf ctrlc_buf; bool ctrlc_flag = false; -void handle_signals(int signo) { +void handle_signals(int signo) +{ if (signo == SIGINT) { ctrlc_flag = true; siglongjmp(ctrlc_buf, 1); @@ -61,7 +62,8 @@ char *my_readline(const char *prompt) } } - while ( sigsetjmp( ctrlc_buf, 1 ) != 0 ); + while (sigsetjmp(ctrlc_buf, 1) != 0) + ; if (ctrlc_flag) { char *line = (char *)malloc(strlen("exit") + 1); @@ -114,10 +116,8 @@ char *my_readline(const char *prompt) */ bool is_exit_command(const char *cmd) { - return 0 == strncasecmp("exit", cmd, 4) - || 0 == strncasecmp("bye", cmd, 3) - || 0 == strncasecmp("\\q", cmd, 2) - || 0 == strncasecmp("interrupted", cmd, 11); + return 0 == strncasecmp("exit", cmd, 4) || 0 == strncasecmp("bye", cmd, 3) || 0 == strncasecmp("\\q", cmd, 2) || + 0 == strncasecmp("interrupted", cmd, 11); } char *read_command() diff --git a/src/observer/net/java_thread_pool_thread_handler.cpp b/src/observer/net/java_thread_pool_thread_handler.cpp index 476fa916..f58a929a 100644 --- a/src/observer/net/java_thread_pool_thread_handler.cpp +++ b/src/observer/net/java_thread_pool_thread_handler.cpp @@ -25,13 +25,13 @@ using namespace common; /** * @brief libevent 消息回调函数的参数 - * + * */ struct EventCallbackAg { - JavaThreadPoolThreadHandler *host = nullptr; - Communicator *communicator = nullptr; - struct event *ev = nullptr; + JavaThreadPoolThreadHandler *host = nullptr; + Communicator *communicator = nullptr; + struct event *ev = nullptr; }; JavaThreadPoolThreadHandler::~JavaThreadPoolThreadHandler() @@ -59,10 +59,10 @@ RC JavaThreadPoolThreadHandler::start() // 创建线程池 // 这里写死了线程池的大小，实际上可以从配置文件中读取 int ret = executor_.init("SQL", // name - 2, // core size - 8, // max size - 60*1000 // keep alive time - ); + 2, // core size + 8, // max size + 60 * 1000 // keep alive time + ); if (0 != ret) { LOG_ERROR("failed to init thread pool executor"); return RC::INTERNAL; @@ -71,7 +71,7 @@ RC JavaThreadPoolThreadHandler::start() // libevent 的监测消息循环主体，要放在一个线程中执行 // event_loop_thread 是运行libevent 消息监测循环的函数，会长期运行，并且会放到线程池中占据一个线程 auto event_worker = std::bind(&JavaThreadPoolThreadHandler::event_loop_thread, this); - ret = executor_.execute(event_worker); + ret = executor_.execute(event_worker); if (0 != ret) { LOG_ERROR("failed to execute event worker"); return RC::INTERNAL; @@ -94,7 +94,7 @@ static void event_callback(evutil_socket_t fd, short event, void *arg) { if (event & (EV_READ | EV_CLOSED)) { LOG_TRACE("got event. fd=%d, event=%d", fd, event); - EventCallbackAg *ag = (EventCallbackAg *)arg; + EventCallbackAg *ag = (EventCallbackAg *)arg; JavaThreadPoolThreadHandler *handler = ag->host; handler->handle_event(ag); } else { @@ -113,7 +113,7 @@ void JavaThreadPoolThreadHandler::handle_event(EventCallbackAg *ag) // sql_handler 是一个回调函数 auto sql_handler = [this, ag]() { - RC rc = sql_task_handler_.handle_event(ag->communicator); // 这里会有接收消息、处理请求然后返回结果一条龙服务 + RC rc = sql_task_handler_.handle_event(ag->communicator); // 这里会有接收消息、处理请求然后返回结果一条龙服务 if (RC::SUCCESS != rc) { LOG_WARN("failed to handle sql task. rc=%s", strrc(rc)); this->close_connection(ag->communicator); @@ -128,7 +128,7 @@ void JavaThreadPoolThreadHandler::handle_event(EventCallbackAg *ag) // LOG_TRACE("add event. fd=%d, communicator=%p", event_get_fd(ag->ev), this); } }; - + executor_.execute(sql_handler); } @@ -148,9 +148,9 @@ RC JavaThreadPoolThreadHandler::new_connection(Communicator *communicator) int fd = communicator->fd(); LOG_INFO("new connection. fd=%d", fd); EventCallbackAg *ag = new EventCallbackAg; - ag->host = this; - ag->communicator = communicator; - ag->ev = nullptr; + ag->host = this; + ag->communicator = communicator; + ag->ev = nullptr; /// 创建一个libevent事件对象。其中EV_READ表示可读事件，就是客户端发消息时会触发事件。 /// EV_ET 表示边缘触发，有消息时只会触发一次，不会重复触发。这个标识在Linux平台上是支持的，但是有些平台不支持。 /// 使用EV_ET边缘触发时需要注意一个问题，就是每次一定要把客户端发来的消息都读取完，直到read返回EAGAIN为止。 @@ -185,7 +185,7 @@ RC JavaThreadPoolThreadHandler::close_connection(Communicator *communicator) { lock_guard guard(lock_); - auto iter = event_map_.find(communicator); + auto iter = event_map_.find(communicator); if (iter == event_map_.end()) { LOG_ERROR("cannot find event for communicator %p", communicator); return RC::INTERNAL; @@ -196,8 +196,8 @@ RC JavaThreadPoolThreadHandler::close_connection(Communicator *communicator) } if (ag->ev) { - event_del(ag->ev); // 把当前事件从event_base中删除 - event_free(ag->ev); // 释放event对象 + event_del(ag->ev); // 把当前事件从event_base中删除 + event_free(ag->ev); // 释放event对象 ag->ev = nullptr; } delete ag; diff --git a/src/observer/net/mysql_communicator.cpp b/src/observer/net/mysql_communicator.cpp index 95dd1521..6936e0bd 100644 --- a/src/observer/net/mysql_communicator.cpp +++ b/src/observer/net/mysql_communicator.cpp @@ -349,7 +349,7 @@ struct HandshakeV10 : public BasePacket char *buf = net_packet.data(); int pos = 0; - pos += 3; // skip packet length + pos += 3; // skip packet length pos += store_int1(buf + pos, packet_header.sequence_id); pos += store_int1(buf + pos, protocol); @@ -380,12 +380,12 @@ struct HandshakeV10 : public BasePacket */ struct OkPacket : public BasePacket { - int8_t header = 0; // 0x00 for ok and 0xFE for EOF - int32_t affected_rows = 0; - int32_t last_insert_id = 0; - int16_t status_flags = 0x22; - int16_t warnings = 0; - string info; // human readable status information + int8_t header = 0; // 0x00 for ok and 0xFE for EOF + int32_t affected_rows = 0; + int32_t last_insert_id = 0; + int16_t status_flags = 0x22; + int16_t warnings = 0; + string info; // human readable status information OkPacket(int8_t sequence = 0) : BasePacket(sequence) {} virtual ~OkPacket() = default; @@ -399,7 +399,7 @@ struct OkPacket : public BasePacket char *buf = net_packet.data(); int pos = 0; - pos += 3; // skip packet length + pos += 3; // skip packet length pos += store_int1(buf + pos, packet_header.sequence_id); pos += store_int1(buf + pos, header); pos += store_lenenc_int(buf + pos, affected_rows); @@ -472,11 +472,11 @@ struct EofPacket : public BasePacket */ struct ErrPacket : public BasePacket { - int8_t header = 0xFF; - int16_t error_code = 0; - char sql_state_marker[1] = {'#'}; - string sql_state{"HY000"}; - string error_message; + int8_t header = 0xFF; + int16_t error_code = 0; + char sql_state_marker[1] = {'#'}; + string sql_state{"HY000"}; + string error_message; ErrPacket(int8_t sequence = 0) : BasePacket(sequence) {} virtual ~ErrPacket() = default; @@ -520,7 +520,7 @@ struct QueryPacket { PacketHeader packet_header; int8_t command; // 0x03: COM_QUERY - string query; // the text of the SQL query to execute + string query; // the text of the SQL query to execute }; /** @@ -677,6 +677,18 @@ RC MysqlCommunicator::read_event(SessionEvent *&event) return handle_version_comment(need_disconnect); } + if (query_packet.query.find("SET NAMES utf8mb4") != string::npos) { + /// 该设置暂时不支持 + OkPacket ok_packet(sequence_id_); + rc = send_packet(ok_packet); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to send ok packet. command=%d, addr=%s, error=%s", command_type, addr(), strrc(rc)); + return rc; + } + writer_->flush(); + return rc; + } + event = new SessionEvent(this); event->set_query(query_packet.query); } else { @@ -696,8 +708,8 @@ RC MysqlCommunicator::write_state(SessionEvent *event, bool &need_disconnect) { SqlResult *sql_result = event->sql_result(); - const int buf_size = 2048; - char *buf = new char[buf_size]; + const int buf_size = 2048; + char *buf = new char[buf_size]; const string &state_string = sql_result->state_string(); if (state_string.empty()) { const char *result = strrc(sql_result->return_code()); @@ -938,16 +950,16 @@ RC MysqlCommunicator::send_column_definition(SqlResult *sql_result, bool &need_d * @param no_column_def 为了特殊处理没有返回值的语句，比如insert/delete，需要做特殊处理。 * 这种语句只需要返回一个ok packet即可 */ -RC MysqlCommunicator::send_result_rows(SessionEvent *event, SqlResult *sql_result, bool no_column_def, bool &need_disconnect) +RC MysqlCommunicator::send_result_rows( + SessionEvent *event, SqlResult *sql_result, bool no_column_def, bool &need_disconnect) { RC rc = RC::SUCCESS; vector packet; packet.resize(4 * 1024 * 1024); // TODO warning: length cannot be fix - int affected_rows = 0; - if (event->session()->get_execution_mode() == ExecutionMode::CHUNK_ITERATOR - && event->session()->used_chunk_mode()) { + int affected_rows = 0; + if (event->session()->get_execution_mode() == ExecutionMode::CHUNK_ITERATOR && event->session()->used_chunk_mode()) { rc = write_chunk_result(sql_result, packet, affected_rows, need_disconnect); } else { rc = write_tuple_result(sql_result, packet, affected_rows, need_disconnect); @@ -972,10 +984,11 @@ RC MysqlCommunicator::send_result_rows(SessionEvent *event, SqlResult *sql_resul return rc; } -RC MysqlCommunicator::write_tuple_result(SqlResult *sql_result, vector &packet, int &affected_rows, bool &need_disconnect) +RC MysqlCommunicator::write_tuple_result( + SqlResult *sql_result, vector &packet, int &affected_rows, bool &need_disconnect) { - Tuple *tuple = nullptr; - RC rc = RC::SUCCESS; + Tuple *tuple = nullptr; + RC rc = RC::SUCCESS; while (RC::SUCCESS == (rc = sql_result->next_tuple(tuple))) { assert(tuple != nullptr); @@ -1017,10 +1030,11 @@ RC MysqlCommunicator::write_tuple_result(SqlResult *sql_result, vector &pa } return rc; } -RC MysqlCommunicator::write_chunk_result(SqlResult *sql_result, vector &packet, int &affected_rows, bool &need_disconnect) +RC MysqlCommunicator::write_chunk_result( + SqlResult *sql_result, vector &packet, int &affected_rows, bool &need_disconnect) { Chunk chunk; - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; while (RC::SUCCESS == (rc = sql_result->next_chunk(chunk))) { int column_num = chunk.column_num(); if (column_num == 0) { diff --git a/src/observer/net/one_thread_per_connection_thread_handler.cpp b/src/observer/net/one_thread_per_connection_thread_handler.cpp index 91ff6904..b02c26d3 100644 --- a/src/observer/net/one_thread_per_connection_thread_handler.cpp +++ b/src/observer/net/one_thread_per_connection_thread_handler.cpp @@ -28,9 +28,7 @@ using namespace common; class Worker { public: - Worker(ThreadHandler &host, Communicator *communicator) - : host_(host), communicator_(communicator) - {} + Worker(ThreadHandler &host, Communicator *communicator) : host_(host), communicator_(communicator) {} ~Worker() { if (thread_ != nullptr) { @@ -55,7 +53,7 @@ class Worker { if (thread_) { if (thread_->get_id() == this_thread::get_id()) { - thread_->detach(); // 如果当前线程join当前线程，就会卡死 + thread_->detach(); // 如果当前线程join当前线程，就会卡死 } else { thread_->join(); } @@ -74,8 +72,8 @@ class Worker } struct pollfd poll_fd; - poll_fd.fd = communicator_->fd(); - poll_fd.events = POLLIN; + poll_fd.fd = communicator_->fd(); + poll_fd.events = POLLIN; poll_fd.revents = 0; while (running_) { @@ -101,15 +99,15 @@ class Worker } LOG_INFO("worker thread stop. communicator = %p", communicator_); - host_.close_connection(communicator_); /// 连接关闭后，当前对象会被删除 + host_.close_connection(communicator_); /// 连接关闭后，当前对象会被删除 } private: ThreadHandler &host_; SqlTaskHandler task_handler_; - Communicator *communicator_ = nullptr; - thread *thread_ = nullptr; - volatile bool running_ = true; + Communicator *communicator_ = nullptr; + thread *thread_ = nullptr; + volatile bool running_ = true; }; OneThreadPerConnectionThreadHandler::~OneThreadPerConnectionThreadHandler() @@ -128,7 +126,7 @@ RC OneThreadPerConnectionThreadHandler::new_connection(Communicator *communicato return RC::FILE_EXIST; } - Worker *worker = new Worker(*this, communicator); + Worker *worker = new Worker(*this, communicator); thread_map_[communicator] = worker; return worker->start(); } diff --git a/src/observer/net/plain_communicator.cpp b/src/observer/net/plain_communicator.cpp index e1674671..621ee8e6 100644 --- a/src/observer/net/plain_communicator.cpp +++ b/src/observer/net/plain_communicator.cpp @@ -37,7 +37,7 @@ RC PlainCommunicator::read_event(SessionEvent *&event) int data_len = 0; int read_len = 0; - const int max_packet_size = 8192; + const int max_packet_size = 131072; vector buf(max_packet_size); // 持续接收消息，直到遇到'\0'。将'\0'遇到的后续数据直接丢弃没有处理，因为目前仅支持一收一发的模式 @@ -160,6 +160,7 @@ RC PlainCommunicator::write_debug(SessionEvent *request, bool &need_disconnect) RC PlainCommunicator::write_result(SessionEvent *event, bool &need_disconnect) { RC rc = write_result_internal(event, need_disconnect); + if (!need_disconnect) { RC rc1 = write_debug(event, need_disconnect); if (OB_FAIL(rc1)) { @@ -180,8 +181,7 @@ RC PlainCommunicator::write_result(SessionEvent *event, bool &need_disconnect) RC PlainCommunicator::write_result_internal(SessionEvent *event, bool &need_disconnect) { - RC rc = RC::SUCCESS; - + RC rc = RC::SUCCESS; need_disconnect = true; SqlResult *sql_result = event->sql_result(); @@ -205,47 +205,58 @@ RC PlainCommunicator::write_result_internal(SessionEvent *event, bool &need_disc const char *alias = spec.alias(); if (nullptr != alias || alias[0] != 0) { if (0 != i) { - const char *delim = " | "; - - rc = writer_->writen(delim, strlen(delim)); - if (OB_FAIL(rc)) { - LOG_WARN("failed to send data to client. err=%s", strerror(errno)); - return rc; - } - } - - int len = strlen(alias); - - rc = writer_->writen(alias, len); - if (OB_FAIL(rc)) { - LOG_WARN("failed to send data to client. err=%s", strerror(errno)); - sql_result->close(); - return rc; + title_stream << " | "; } + title_stream << alias; // 将alias存入流中 } } if (cell_num > 0) { - char newline = '\n'; - - rc = writer_->writen(&newline, 1); - if (OB_FAIL(rc)) { - LOG_WARN("failed to send data to client. err=%s", strerror(errno)); - sql_result->close(); - return rc; - } + title_stream << '\n'; // 将换行符存入流中 } - rc = RC::SUCCESS; - if (event->session()->get_execution_mode() == ExecutionMode::CHUNK_ITERATOR - && event->session()->used_chunk_mode()) { + if (event->session()->get_execution_mode() == ExecutionMode::CHUNK_ITERATOR && event->session()->used_chunk_mode()) { rc = write_chunk_result(sql_result); } else { rc = write_tuple_result(sql_result); } if (OB_FAIL(rc)) { - return rc; + // 清空输出流，避免下次查询输出上次查询失败未输出的内容 + title_stream.str(""); // 清空流内容 + title_stream.clear(); // 重置流状态标志 + sql_result_stream.str(""); // 清空流内容 + sql_result_stream.clear(); // 重置流状态标志 + sql_result->close(); + sql_result->set_return_code(rc); + return write_state(event, need_disconnect); + } else { + // 将 title_stream 中的内容一次性写入到 writer_ + std::string buffer = title_stream.str(); // 获取整个缓冲区的内容 + title_stream.str(""); // 清空流内容 + title_stream.clear(); // 重置流状态（但不会清空内容） + + rc = writer_->writen(buffer.c_str(), buffer.size()); + if (OB_FAIL(rc)) { + LOG_WARN("failed to send data to client. err=%s", strerror(errno)); + sql_result->close(); + return rc; + } + + buffer.clear(); // 清空字符串内容 + + // 将 sql_result_stream 中的内容一次性写入到 writer_ + buffer = sql_result_stream.str(); // 获取缓冲区的内容 + sql_result_stream.str(""); // 清空流内容 + sql_result_stream.clear(); // 重置流状态 + + rc = writer_->writen(buffer.c_str(), buffer.size()); + buffer.clear(); // 清空字符串内容 + if (OB_FAIL(rc)) { + LOG_WARN("failed to send data to client. err=%s", strerror(errno)); + sql_result->close(); + return rc; + } } if (cell_num == 0) { @@ -272,8 +283,9 @@ RC PlainCommunicator::write_result_internal(SessionEvent *event, bool &need_disc RC PlainCommunicator::write_tuple_result(SqlResult *sql_result) { - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; Tuple *tuple = nullptr; + while (RC::SUCCESS == (rc = sql_result->next_tuple(tuple))) { assert(tuple != nullptr); @@ -281,13 +293,7 @@ RC PlainCommunicator::write_tuple_result(SqlResult *sql_result) for (int i = 0; i < cell_num; i++) { if (i != 0) { const char *delim = " | "; - - rc = writer_->writen(delim, strlen(delim)); - if (OB_FAIL(rc)) { - LOG_WARN("failed to send data to client. err=%s", strerror(errno)); - sql_result->close(); - return rc; - } + sql_result_stream << delim; // 将分隔符存入流中 } Value value; @@ -298,24 +304,13 @@ RC PlainCommunicator::write_tuple_result(SqlResult *sql_result) return rc; } - string cell_str = value.to_string(); - - rc = writer_->writen(cell_str.data(), cell_str.size()); - if (OB_FAIL(rc)) { - LOG_WARN("failed to send data to client. err=%s", strerror(errno)); - sql_result->close(); - return rc; - } + // 将cell的值存入流中 + std::string cell_str = value.to_string(); + sql_result_stream << cell_str; } - char newline = '\n'; - - rc = writer_->writen(&newline, 1); - if (OB_FAIL(rc)) { - LOG_WARN("failed to send data to client. err=%s", strerror(errno)); - sql_result->close(); - return rc; - } + // 将换行符存入流中 + sql_result_stream << '\n'; } if (rc == RC::RECORD_EOF) { @@ -326,7 +321,7 @@ RC PlainCommunicator::write_tuple_result(SqlResult *sql_result) RC PlainCommunicator::write_chunk_result(SqlResult *sql_result) { - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; Chunk chunk; while (RC::SUCCESS == (rc = sql_result->next_chunk(chunk))) { int col_num = chunk.column_num(); @@ -370,4 +365,4 @@ RC PlainCommunicator::write_chunk_result(SqlResult *sql_result) rc = RC::SUCCESS; } return rc; -} \ No newline at end of file +} diff --git a/src/observer/net/plain_communicator.h b/src/observer/net/plain_communicator.h index 0dfabf9a..014fe8fe 100644 --- a/src/observer/net/plain_communicator.h +++ b/src/observer/net/plain_communicator.h @@ -41,6 +41,8 @@ class PlainCommunicator : public Communicator RC write_chunk_result(SqlResult *sql_result); protected: - vector send_message_delimiter_; ///< 发送消息分隔符 - vector debug_message_prefix_; ///< 调试信息前缀 + vector send_message_delimiter_; ///< 发送消息分隔符 + vector debug_message_prefix_; ///< 调试信息前缀 + std::ostringstream title_stream; + std::ostringstream sql_result_stream; }; diff --git a/src/observer/net/ring_buffer.cpp b/src/observer/net/ring_buffer.cpp index c3b5f74e..dfd28182 100644 --- a/src/observer/net/ring_buffer.cpp +++ b/src/observer/net/ring_buffer.cpp @@ -32,7 +32,7 @@ RC RingBuffer::read(char *buf, int32_t size, int32_t &read_size) RC rc = RC::SUCCESS; read_size = 0; - while (OB_SUCC(rc) && read_sizesize()> 0) { + while (OB_SUCC(rc) && read_size < size && this->size() > 0) { const char *tmp_buf = nullptr; int32_t tmp_size = 0; rc = buffer(tmp_buf, tmp_size); @@ -90,7 +90,7 @@ RC RingBuffer::write(const char *data, int32_t size, int32_t &write_size) RC rc = RC::SUCCESS; write_size = 0; - while (OB_SUCC(rc) && write_sizeremain()> 0) { + while (OB_SUCC(rc) && write_size < size && this->remain() > 0) { const int32_t read_pos = this->read_pos(); const int32_t tmp_buf_size = (read_pos <= write_pos_) ? (capacity() - write_pos_) : (read_pos - write_pos_); diff --git a/src/observer/net/sql_task_handler.cpp b/src/observer/net/sql_task_handler.cpp index 90f0609b..18b07415 100644 --- a/src/observer/net/sql_task_handler.cpp +++ b/src/observer/net/sql_task_handler.cpp @@ -21,7 +21,7 @@ See the Mulan PSL v2 for more details. */ RC SqlTaskHandler::handle_event(Communicator *communicator) { SessionEvent *event = nullptr; - RC rc = communicator->read_event(event); + RC rc = communicator->read_event(event); if (OB_FAIL(rc)) { return rc; } @@ -88,4 +88,4 @@ RC SqlTaskHandler::handle_sql(SQLStageEvent *sql_event) } return rc; -} \ No newline at end of file +} diff --git a/src/observer/net/thread_handler.cpp b/src/observer/net/thread_handler.cpp index 941fcdfd..b9ae5155 100644 --- a/src/observer/net/thread_handler.cpp +++ b/src/observer/net/thread_handler.cpp @@ -20,7 +20,7 @@ See the Mulan PSL v2 for more details. */ #include "common/log/log.h" #include "common/lang/string.h" -ThreadHandler * ThreadHandler::create(const char *name) +ThreadHandler *ThreadHandler::create(const char *name) { const char *default_name = "one-thread-per-connection"; if (nullptr == name || common::is_blank(name)) { diff --git a/src/observer/sql/builtin/builtin.cpp b/src/observer/sql/builtin/builtin.cpp new file mode 100644 index 00000000..00526696 --- /dev/null +++ b/src/observer/sql/builtin/builtin.cpp @@ -0,0 +1,476 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +#include "builtin.h" + +namespace builtin { + +RC _typeof(const vector &args, Value &result) +{ + if (args.size() != 1) { + return RC::INVALID_ARGUMENT; + } + result = Value(attr_type_to_string(args[0].attr_type())); + return RC::SUCCESS; +} + +RC length(const vector &args, Value &result) +{ + if (args.size() != 1) { + return RC::INVALID_ARGUMENT; + } + if (args[0].attr_type() != AttrType::CHARS) { + return RC::INVALID_ARGUMENT; + } + int length = static_cast(args[0].to_string().size()); + result = Value(length); + return RC::SUCCESS; +} + +RC round(const vector &args, Value &result) +{ + if (args.size() != 1 && args.size() != 2) { + return RC::INVALID_ARGUMENT; + } + float number; + int decimals = 0; // 默认四舍五入为整数 + if (args[0].attr_type() != AttrType::FLOATS) { + return RC::INVALID_ARGUMENT; + } + if (args.size() == 2) { + if (args[1].attr_type() != AttrType::INTS) { + return RC::INVALID_ARGUMENT; + } + decimals = args[1].get_int(); + } + number = args[0].get_float(); + + double round; + double factor = std::pow(10.0, decimals); + double scaledNumber = number * factor; + + // 获取整数部分和小数部分 + double integer_part; + double fractional_part = std::modf(scaledNumber, &integer_part); + + // 如果小数部分刚好是 0.5，进行银行家舍入 + if (fractional_part == 0.5 || fractional_part == -0.5) { + if (static_cast(integer_part) % 2 == 0) { + round = integer_part / factor; // 偶数，直接舍去小数 + } else { + round = (integer_part + (number > 0 ? 1 : -1)) / factor; // 奇数，舍入到偶数 + } + } else { + round = std::round(scaledNumber) / factor; // 否则使用普通的四舍五入 + } + + result = Value(static_cast(round)); + return RC::SUCCESS; +} + +namespace date { +static string get_day_with_suffix(int day) +{ + if (day >= 11 && day <= 13) { + return std::to_string(day) + "th"; + } + switch (day % 10) { + case 1: { + return std::to_string(day) + "st"; + } + case 2: { + return std::to_string(day) + "nd"; + } + case 3: { + return std::to_string(day) + "rd"; + } + default: { + return std::to_string(day) + "th"; + } + } +} + +static string get_full_month_name(int month) +{ + switch (month) { + case 1: return "January"; + case 2: return "February"; + case 3: return "March"; + case 4: return "April"; + case 5: return "May"; + case 6: return "June"; + case 7: return "July"; + case 8: return "August"; + case 9: return "September"; + case 10: return "October"; + case 11: return "November"; + case 12: return "December"; + default: return ""; // 如果月份值无效，返回一个错误字符串 + } +} + +static RC get_year_month_day(const Value &value, int &year, int &month, int &day) +{ + if (value.attr_type() == AttrType::DATES) { + // 提取年、月、日（假设日期格式为YYYYMMDD） + int val = value.get_date(); + year = val / 10000; // 获取年份 + month = (val / 100) % 100; // 获取月份 + day = val % 100; // 获取日期 + } + + if (value.attr_type() == AttrType::CHARS) { + // 日期格式假设为 '2019-9-17' 或 '2019-09-17' + std::string date_str = value.to_string(); + if (sscanf(date_str.c_str(), "%d-%d-%d", &year, &month, &day) != 3) { + return RC::INVALID_ARGUMENT; + } + } + + if (!check_date(year, month, day)) { + return RC::ERROR_DATE; + } + + return RC::SUCCESS; +} + +} // namespace date + +RC year(const vector &args, Value &result) +{ + if (args.size() != 1) { + return RC::INVALID_ARGUMENT; + } + if (args[0].attr_type() != AttrType::DATES && args[0].attr_type() != AttrType::CHARS) { + return RC::INVALID_ARGUMENT; + } + int year, month, day; + RC rc = date::get_year_month_day(args[0], year, month, day); + if (OB_FAIL(rc)) { + return rc; + } + result = Value(year); + return RC::SUCCESS; +} + +RC month(const vector &args, Value &result) +{ + if (args.size() != 1) { + return RC::INVALID_ARGUMENT; + } + if (args[0].attr_type() != AttrType::DATES && args[0].attr_type() != AttrType::CHARS) { + return RC::INVALID_ARGUMENT; + } + int year, month, day; + RC rc = date::get_year_month_day(args[0], year, month, day); + if (OB_FAIL(rc)) { + return rc; + } + result = Value(month); + return RC::SUCCESS; +} + +RC day(const vector &args, Value &result) +{ + if (args.size() != 1) { + return RC::INVALID_ARGUMENT; + } + if (args[0].attr_type() != AttrType::DATES && args[0].attr_type() != AttrType::CHARS) { + return RC::INVALID_ARGUMENT; + } + int year, month, day; + RC rc = date::get_year_month_day(args[0], year, month, day); + if (OB_FAIL(rc)) { + return rc; + } + result = Value(day); + return RC::SUCCESS; +} + +RC date_format(const vector &args, Value &result) +{ + if (args.size() != 2) { + return RC::INVALID_ARGUMENT; + } + if (args[0].attr_type() != AttrType::DATES && args[0].attr_type() != AttrType::CHARS) { + return RC::INVALID_ARGUMENT; + } + if (args[1].attr_type() != AttrType::CHARS) { + return RC::INVALID_ARGUMENT; + } + + int year, month, day; + RC rc = date::get_year_month_day(args[0], year, month, day); + if (OB_FAIL(rc)) { + return rc; + } + + auto fmt = args[1].to_string(); + + string str; + + // 遍历格式字符串，并替换格式符 + for (size_t i = 0; i < fmt.length(); ++i) { + if (fmt[i] == '%' && i + 1 < fmt.length()) { + switch (fmt[i + 1]) { + case 'Y': // 四位数年份 + str += std::to_string(year); + break; + case 'y': // 两位数年份 + str += std::to_string(year).substr(2, 2); + break; + case 'm': // 两位数月份 + str += (month < 10 ? "0" : "") + std::to_string(month); + break; + case 'c': // 不带前导零的月份 + str += std::to_string(month); + break; + case 'M': // 完整的月份名称 + str += date::get_full_month_name(month); + break; + case 'd': // 两位数日期 + str += (day < 10 ? "0" : "") + std::to_string(day); + break; + case 'e': // 不带前导零的日期 + str += std::to_string(day); + break; + case 'D': // 带序数后缀的日期 + str += date::get_day_with_suffix(day); + break; + default: // 未知格式符，按原样输出 + str += fmt[i + 1]; + break; + } + ++i; // 跳过格式符的下一个字符 + } else { + str += fmt[i]; // 普通字符直接追加 + } + } + + result = Value(str.c_str()); + return RC::SUCCESS; +} + +namespace vector_distance { + +RC distance(const std::vector &args, Value &result, NormalFunctionType type) +{ + if (args.size() != 2) { + return RC::INVALID_ARGUMENT; + } + if (args[0].attr_type() != AttrType::VECTORS && args[0].attr_type() != AttrType::CHARS) { + return RC::INVALID_ARGUMENT; + } + if (args[1].attr_type() != AttrType::VECTORS && args[1].attr_type() != AttrType::CHARS) { + return RC::INVALID_ARGUMENT; + } + + Value value0, value1; + if (args[0].attr_type() == AttrType::CHARS) { + RC rc = Value::cast_to(args[0], AttrType::VECTORS, value0); + if (OB_FAIL(rc)) { + return rc; + } + } else if (args[0].attr_type() == AttrType::VECTORS) { + value0 = args[0]; + } else { + return RC::INVALID_ARGUMENT; + } + + if (args[1].attr_type() == AttrType::CHARS) { + RC rc = Value::cast_to(args[1], AttrType::VECTORS, value1); + if (OB_FAIL(rc)) { + return rc; + } + } else if (args[1].attr_type() == AttrType::VECTORS) { + value1 = args[1]; + } else { + return RC::INVALID_ARGUMENT; + } + + auto v0_length = value0.get_vector_length(); + auto v1_length = value1.get_vector_length(); + if (v0_length != v1_length) { + return RC::VECTOR_LENGTG_ARE_INCONSISTENT; + } + + switch (type) { + case NormalFunctionType::L2_DISTANCE: { + /* + * l2_distance + * 语法：l2_distance(vector A, vector B) + * 计算公式：$[ D = \sqrt{\sum_{i=1}^{n} (A_{i} - B_{i})^2} ]$ + */ + float ans = 0.0; + for (int i = 0; i < v0_length; i++) { + float v0 = args[0].get_vector_element(i); + float v1 = args[1].get_vector_element(i); + ans += (v0 - v1) * (v0 - v1); + } + ans = sqrt(ans); + result = Value(ans); + return RC::SUCCESS; + } + case NormalFunctionType::COSINE_DISTANCE: { + /* + * cosine_distance： + * 语法：cosine_distance(vector A, vector B) + * 计算公式：$[ D = \frac{\mathbf{A} \cdot \mathbf{B}}{|\mathbf{A}| |\mathbf{B}|} = \frac{\sum_{i=1}^{n} A_i * + * B_i}{\sqrt{\sum_{i=1}^{n} A_i^2} \sqrt{\sum_{i=1}^{n} B_i^2}} ]$ + */ + float dot_product = 0.0; + float norm_v0 = 0.0; + float norm_v1 = 0.0; + + for (int i = 0; i < v0_length; i++) { + float v0 = args[0].get_vector_element(i); + float v1 = args[1].get_vector_element(i); + dot_product += v0 * v1; // 计算点积 + norm_v0 += v0 * v0; // 计算 v0 的模长平方 + norm_v1 += v1 * v1; // 计算 v1 的模长平方 + } + + if (std::fabs(norm_v0) < EPSILON || std::fabs(norm_v1) < EPSILON) { + result = Value(NullValue()); + return RC::SUCCESS; // 避免除以 0 的情况 + } + + float cosine_similarity = dot_product / (std::sqrt(norm_v0) * std::sqrt(norm_v1)); + result = Value(1 - cosine_similarity); + return RC::SUCCESS; + } + case NormalFunctionType::INNER_PRODUCT: { + /* + * inner_product： + * 语法：inner_product(vector A, vector B) + * 计算公式：$[ D = \mathbf{A} \cdot \mathbf{B} = a_1 b_1 + a_2 b_2 + ... + a_n b_n = \sum_{i=1}^{n} a_i b_i ]$ + */ + float dot_product = 0.0; + + for (int i = 0; i < v0_length; i++) { + float v0 = args[0].get_vector_element(i); + float v1 = args[1].get_vector_element(i); + dot_product += v0 * v1; // 对应元素相乘并求和 + } + + result = Value(dot_product); + return RC::SUCCESS; + } + default: { + return RC::INTERNAL; + } + } +} +} // namespace vector_distance + +RC l2_distance(const vector &args, Value &result) +{ + return vector_distance::distance(args, result, NormalFunctionType::L2_DISTANCE); +} + +RC cosine_distance(const vector &args, Value &result) +{ + return vector_distance::distance(args, result, NormalFunctionType::COSINE_DISTANCE); +} + +RC inner_product(const vector &args, Value &result) +{ + return vector_distance::distance(args, result, NormalFunctionType::INNER_PRODUCT); +} + +// 向量索引调用以下方法 +// L2 距离（欧氏距离） +float l2_distance(const std::vector &a, const std::vector &b) +{ + float sum = 0.0; + for (size_t i = 0; i < a.size(); ++i) { + float diff = a[i] - b[i]; + sum += diff * diff; + } + return std::sqrt(sum); +} + +// 余弦距离 +float cosine_distance(const std::vector &a, const std::vector &b) +{ + float dotProduct = 0.0; + float normASquared = 0.0; + float normBSquared = 0.0; + + for (size_t i = 0; i < a.size(); ++i) { + dotProduct += a[i] * b[i]; + normASquared += a[i] * a[i]; + normBSquared += b[i] * b[i]; + } + + float denom = std::sqrt(normASquared) * std::sqrt(normBSquared); + if (denom == 0.0) { + return 1.0; // 避免除以 0 的情况 + } + + float cosineSimilarity = dotProduct / denom; + return 1.0 - cosineSimilarity; // 将相似度转变为距离 +} + +// 内积 +float inner_product(const std::vector &a, const std::vector &b) +{ + float product = 0.0; + for (size_t i = 0; i < a.size(); ++i) { + product += a[i] * b[i]; + } + return product; +} + +RC string_to_vector(const vector &args, Value &result) +{ + if (args.size() != 1) { + return RC::INVALID_ARGUMENT; + } + if (args[0].attr_type() != AttrType::CHARS) { + return RC::INVALID_ARGUMENT; + } + return Value::cast_to(args[0], AttrType::VECTORS, result); +} + +RC vector_to_string(const vector &args, Value &result) +{ + if (args.size() != 1) { + return RC::INVALID_ARGUMENT; + } + if (args[0].attr_type() != AttrType::VECTORS) { + return RC::INVALID_ARGUMENT; + } + result = Value(args[0].to_string().c_str()); + return RC::SUCCESS; +} + +RC vector_dim(const vector &args, Value &result) +{ + if (args.size() != 1) { + return RC::INVALID_ARGUMENT; + } + Value value; + if (args[0].attr_type() == AttrType::CHARS) { + RC rc = Value::cast_to(args[0], AttrType::VECTORS, value); + if (OB_FAIL(rc)) { + return rc; + } + } else if (args[0].attr_type() == AttrType::VECTORS) { + value = args[0]; + } else { + return RC::INVALID_ARGUMENT; + } + + result = Value(value.get_vector_length()); + return RC::SUCCESS; +} + +} // namespace builtin diff --git a/src/observer/sql/builtin/builtin.h b/src/observer/sql/builtin/builtin.h new file mode 100644 index 00000000..c71c9791 --- /dev/null +++ b/src/observer/sql/builtin/builtin.h @@ -0,0 +1,77 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +#pragma once + +#include "common/value.h" +#include "common/utils.h" + +enum AggregateFunctionType +{ + COUNT, + SUM, + AVG, + MAX, + MIN, +}; + +enum class NormalFunctionType +{ + TYPEOF, + LENGTH, + ROUND, + YEAR, + MONTH, + DAY, + DATE_FORMAT, + L2_DISTANCE, + COSINE_DISTANCE, + INNER_PRODUCT, + STRING_TO_VECTOR, + VECTOR_TO_STRING, + VECTOR_DIM, +}; + +namespace builtin { + +extern RC _typeof(const vector &args, Value &result); + +extern RC length(const vector &args, Value &result); + +extern RC round(const vector &args, Value &result); + +extern RC year(const vector &args, Value &result); + +extern RC month(const vector &args, Value &result); + +extern RC day(const vector &args, Value &result); + +extern RC date_format(const vector &args, Value &result); + +extern RC l2_distance(const vector &args, Value &result); + +extern RC cosine_distance(const vector &args, Value &result); + +extern RC inner_product(const vector &args, Value &result); + +// 以下方法为向量索引中使用 +extern float l2_distance(const std::vector &a, const std::vector &b); + +extern float cosine_distance(const std::vector &a, const std::vector &b); + +extern float inner_product(const std::vector &a, const std::vector &b); + +extern RC string_to_vector(const vector &args, Value &result); + +extern RC vector_to_string(const vector &args, Value &result); + +extern RC vector_dim(const vector &args, Value &result); + +}; // namespace builtin diff --git a/src/observer/sql/executor/command_executor.cpp b/src/observer/sql/executor/command_executor.cpp index 0da5c4e0..35b603c9 100644 --- a/src/observer/sql/executor/command_executor.cpp +++ b/src/observer/sql/executor/command_executor.cpp @@ -13,10 +13,14 @@ See the Mulan PSL v2 for more details. */ // #include "sql/executor/command_executor.h" +#include "sql/executor/create_view_executor.h" +#include "sql/executor/drop_view_executor.h" +#include "sql/executor/show_index_executor.h" #include "common/log/log.h" #include "event/sql_event.h" #include "sql/executor/create_index_executor.h" #include "sql/executor/create_table_executor.h" +#include "sql/executor/drop_table_executor.h" #include "sql/executor/desc_table_executor.h" #include "sql/executor/help_executor.h" #include "sql/executor/load_data_executor.h" @@ -42,6 +46,21 @@ RC CommandExecutor::execute(SQLStageEvent *sql_event) rc = executor.execute(sql_event); } break; + case StmtType::DROP_TABLE: { + DropTableExecutor executor; + rc = executor.execute(sql_event); + } break; + + case StmtType::CREATE_VIEW: { + CreateViewExecutor executor; + rc = executor.execute(sql_event); + } break; + + case StmtType::DROP_VIEW: { + DropViewExecutor executor; + rc = executor.execute(sql_event); + } break; + case StmtType::DESC_TABLE: { DescTableExecutor executor; rc = executor.execute(sql_event); @@ -57,6 +76,11 @@ RC CommandExecutor::execute(SQLStageEvent *sql_event) rc = executor.execute(sql_event); } break; + case StmtType::SHOW_INDEX: { + ShowIndexExecutor executor; + rc = executor.execute(sql_event); + } break; + case StmtType::BEGIN: { TrxBeginExecutor executor; rc = executor.execute(sql_event); diff --git a/src/observer/sql/executor/create_index_executor.cpp b/src/observer/sql/executor/create_index_executor.cpp index 6c99dda0..a955c473 100644 --- a/src/observer/sql/executor/create_index_executor.cpp +++ b/src/observer/sql/executor/create_index_executor.cpp @@ -32,5 +32,21 @@ RC CreateIndexExecutor::execute(SQLStageEvent *sql_event) Trx *trx = session->current_trx(); Table *table = create_index_stmt->table(); - return table->create_index(trx, create_index_stmt->field_meta(), create_index_stmt->index_name().c_str()); -} \ No newline at end of file + + if (create_index_stmt->index_type() == IndexType::BPlusTreeIndex) { + return table->create_index(trx, + create_index_stmt->index_type(), + create_index_stmt->field_meta(), + create_index_stmt->index_name().c_str(), + create_index_stmt->unique()); + } + if (create_index_stmt->index_type() == IndexType::VectorIVFFlatIndex) { + return table->create_vector_index(trx, + create_index_stmt->index_type(), + create_index_stmt->field_meta(), + create_index_stmt->index_name().c_str(), + create_index_stmt->distance_type(), + create_index_stmt->options()); + } + return RC::UNSUPPORTED; +} diff --git a/src/observer/sql/executor/create_table_executor.cpp b/src/observer/sql/executor/create_table_executor.cpp index 3ef9cac9..4d20352a 100644 --- a/src/observer/sql/executor/create_table_executor.cpp +++ b/src/observer/sql/executor/create_table_executor.cpp @@ -20,9 +20,15 @@ See the Mulan PSL v2 for more details. */ #include "session/session.h" #include "sql/stmt/create_table_stmt.h" #include "storage/db/db.h" +#include "sql/optimizer/logical_plan_generator.h" +#include "sql/operator/logical_operator.h" +#include "sql/operator/project_logical_operator.h" +#include "sql/optimizer/physical_plan_generator.h" +#include "storage/trx/trx.h" RC CreateTableExecutor::execute(SQLStageEvent *sql_event) { + RC rc; Stmt *stmt = sql_event->stmt(); Session *session = sql_event->session_event()->session(); ASSERT(stmt->type() == StmtType::CREATE_TABLE, @@ -30,9 +36,85 @@ RC CreateTableExecutor::execute(SQLStageEvent *sql_event) static_cast(stmt->type())); CreateTableStmt *create_table_stmt = static_cast(stmt); + const char *table_name = create_table_stmt->table_name().c_str(); + if (create_table_stmt->create_table_select_stmt()) { + SelectStmt *select_stmt = create_table_stmt->create_table_select_stmt(); - const char *table_name = create_table_stmt->table_name().c_str(); - RC rc = session->get_current_db()->create_table(table_name, create_table_stmt->attr_infos(), create_table_stmt->storage_format()); + std::vector attr_infos = create_table_stmt->attr_infos(); + if (attr_infos.empty()) { + for (auto &expr : select_stmt->query_expressions()) { + AttrInfoSqlNode attr_info; + if (FieldExpr *field_expr = dynamic_cast(expr.get())) { + auto field_meta = field_expr->field().meta(); + attr_info.name = field_meta->name(); + attr_info.type = field_meta->type(); + attr_info.length = field_meta->len(); + attr_info.nullable = field_meta->nullable(); + } else { + attr_info.name = expr->name(); + attr_info.length = expr->value_length() + 1; + attr_info.type = expr->value_type(); + attr_info.nullable = true; + } + attr_info.mutable_ = true; + attr_infos.push_back(attr_info); + } + } + + rc = session->get_current_db()->create_table(table_name, attr_infos, create_table_stmt->storage_format()); + if (OB_FAIL(rc)) { + return rc; + } + + unique_ptr logical_oper = nullptr; + LogicalPlanGenerator::create(create_table_stmt->create_table_select_stmt(), logical_oper); + if (!logical_oper) { + return RC::INTERNAL; + } + + unique_ptr physical_oper = nullptr; + PhysicalPlanGenerator::create(*logical_oper, physical_oper); + auto table_ = session->get_current_db()->find_table(table_name); + if (!physical_oper) { + return RC::INTERNAL; + } + + physical_oper->open(session->current_trx()); + while (RC::SUCCESS == (rc = physical_oper->next())) { + auto tuple = physical_oper->current_tuple(); + int num = tuple->cell_num(); + vector values; + for (int i = 0; i < num; i++) { + Value cell; + rc = tuple->cell_at(i, cell); + if (OB_FAIL(rc)) { + return rc; + } + values.push_back(cell); + } + + Record record; + rc = table_->make_record(static_cast(values.size()), values.data(), record); + if (OB_FAIL(rc)) { + return rc; + } + + rc = session->current_trx()->insert_record(table_, record); + if (OB_FAIL(rc)) { + return rc; + } + } + + rc = physical_oper->close(); + if (OB_FAIL(rc)) { + return rc; + } + + rc = RC::SUCCESS; + } else { + rc = session->get_current_db()->create_table( + table_name, create_table_stmt->attr_infos(), create_table_stmt->storage_format()); + } return rc; -} \ No newline at end of file +} diff --git a/src/observer/sql/executor/create_view_executor.cpp b/src/observer/sql/executor/create_view_executor.cpp new file mode 100644 index 00000000..4ffaca07 --- /dev/null +++ b/src/observer/sql/executor/create_view_executor.cpp @@ -0,0 +1,72 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : create view executor source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "sql/executor/create_view_executor.h" +#include "common/log/log.h" +#include "event/session_event.h" +#include "event/sql_event.h" +#include "session/session.h" +#include "sql/stmt/create_table_stmt.h" +#include "storage/db/db.h" +#include "storage/trx/trx.h" +#include "sql/stmt/create_view_stmt.h" + +// 提取 AS 后的 SQL 语句 +std::string extract_select_sql(const std::string &createViewSql) +{ + std::string lowerSql = createViewSql; + common::str_to_lower(lowerSql); + + std::string asToken = " as "; + + auto pos = lowerSql.find(asToken); + if (pos == std::string::npos) { + return ""; + } + + // 计算实际 SQL 句子的起始位置 + auto actualPos = pos + asToken.length(); + return createViewSql.substr(actualPos); +} + +RC CreateViewExecutor::execute(SQLStageEvent *sql_event) +{ + RC rc; + Stmt *stmt = sql_event->stmt(); + Session *session = sql_event->session_event()->session(); + ASSERT(stmt->type() == StmtType::CREATE_VIEW, + "create view executor can not run this command: %d", + static_cast(stmt->type())); + + auto create_view_stmt = static_cast(stmt); + auto table_name = create_view_stmt->table_name().c_str(); + auto select_stmt = create_view_stmt->select_stmt(); + ASSERT(select_stmt != nullptr, + "create view executor can not run this command: %d", + static_cast(stmt->type())); + + auto select_sql = extract_select_sql(sql_event->sql()); + if (select_sql.empty()) { + return RC::SQL_SYNTAX; + } + + rc = session->get_current_db()->create_table(table_name, + std::move(create_view_stmt->attr_names()), + std::move(select_sql), + select_stmt, + StorageFormat::ROW_FORMAT); + if (OB_FAIL(rc)) { + return rc; + } + + return rc; +} diff --git a/src/observer/sql/executor/create_view_executor.h b/src/observer/sql/executor/create_view_executor.h new file mode 100644 index 00000000..729a5730 --- /dev/null +++ b/src/observer/sql/executor/create_view_executor.h @@ -0,0 +1,30 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : create view executor header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "common/rc.h" + +class SQLStageEvent; + +/** + * @brief 创建视图的执行器 + * @ingroup Executor + */ +class CreateViewExecutor +{ +public: + CreateViewExecutor() = default; + virtual ~CreateViewExecutor() = default; + + RC execute(SQLStageEvent *sql_event); +}; diff --git a/src/observer/sql/executor/desc_table_executor.cpp b/src/observer/sql/executor/desc_table_executor.cpp index 3c3ba0cc..ac2001be 100644 --- a/src/observer/sql/executor/desc_table_executor.cpp +++ b/src/observer/sql/executor/desc_table_executor.cpp @@ -41,8 +41,8 @@ RC DescTableExecutor::execute(SQLStageEvent *sql_event) SqlResult *sql_result = session_event->sql_result(); const char *table_name = desc_table_stmt->table_name().c_str(); - Db *db = session->get_current_db(); - Table *table = db->find_table(table_name); + Db *db = session->get_current_db(); + BaseTable *table = db->find_table(table_name); if (table != nullptr) { TupleSchema tuple_schema; tuple_schema.append_cell(TupleCellSpec("", "Field", "Field")); @@ -65,4 +65,4 @@ RC DescTableExecutor::execute(SQLStageEvent *sql_event) sql_result->set_state_string("Table not exists"); } return rc; -} \ No newline at end of file +} diff --git a/src/observer/sql/executor/drop_table_executor.cpp b/src/observer/sql/executor/drop_table_executor.cpp new file mode 100644 index 00000000..ad5b8398 --- /dev/null +++ b/src/observer/sql/executor/drop_table_executor.cpp @@ -0,0 +1,36 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/15 * + * @Description : DropTableExecutor source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "sql/executor/drop_table_executor.h" + +#include "common/log/log.h" +#include "event/session_event.h" +#include "event/sql_event.h" +#include "session/session.h" +#include "sql/stmt/drop_table_stmt.h" +#include "storage/db/db.h" + +RC DropTableExecutor::execute(SQLStageEvent *sql_event) +{ + Stmt *stmt = sql_event->stmt(); + Session *session = sql_event->session_event()->session(); + ASSERT(stmt->type() == StmtType::DROP_TABLE, + "drop table executor can not run this command: %d", + static_cast(stmt->type())); + + auto *drop_table_stmt = dynamic_cast(stmt); + + const char *table_name = drop_table_stmt->table_name().c_str(); + RC rc = session->get_current_db()->drop_table(table_name); + + return rc; +} diff --git a/src/observer/sql/executor/drop_table_executor.h b/src/observer/sql/executor/drop_table_executor.h new file mode 100644 index 00000000..256b60e5 --- /dev/null +++ b/src/observer/sql/executor/drop_table_executor.h @@ -0,0 +1,30 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/15 * + * @Description : DropTableExecutor header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "common/rc.h" + +class SQLStageEvent; + +/** + * @brief 删除表的执行器 + * @ingroup Executor + */ +class DropTableExecutor +{ +public: + DropTableExecutor() = default; + virtual ~DropTableExecutor() = default; + + RC execute(SQLStageEvent *sql_event); +}; diff --git a/src/observer/sql/executor/drop_view_executor.cpp b/src/observer/sql/executor/drop_view_executor.cpp new file mode 100644 index 00000000..0500c2ec --- /dev/null +++ b/src/observer/sql/executor/drop_view_executor.cpp @@ -0,0 +1,35 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "common/log/log.h" +#include "event/session_event.h" +#include "event/sql_event.h" +#include "session/session.h" +#include "storage/db/db.h" +#include "sql/stmt/drop_view_stmt.h" +#include "sql/executor/drop_view_executor.h" + +RC DropViewExecutor::execute(SQLStageEvent *sql_event) +{ + Stmt *stmt = sql_event->stmt(); + Session *session = sql_event->session_event()->session(); + ASSERT(stmt->type() == StmtType::DROP_VIEW, + "drop view executor can not run this command: %d", + static_cast(stmt->type())); + + auto *drop_view_stmt = dynamic_cast(stmt); + + const char *table_name = drop_view_stmt->table_name().c_str(); + RC rc = session->get_current_db()->drop_table(table_name); + + return rc; +} diff --git a/src/observer/sql/executor/drop_view_executor.h b/src/observer/sql/executor/drop_view_executor.h new file mode 100644 index 00000000..21b5fe7c --- /dev/null +++ b/src/observer/sql/executor/drop_view_executor.h @@ -0,0 +1,30 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "common/rc.h" + +class SQLStageEvent; + +/** + * @brief 删除视图的执行器 + * @ingroup Executor + */ +class DropViewExecutor +{ +public: + DropViewExecutor() = default; + virtual ~DropViewExecutor() = default; + + RC execute(SQLStageEvent *sql_event); +}; diff --git a/src/observer/sql/executor/load_data_executor.cpp b/src/observer/sql/executor/load_data_executor.cpp index c3d6fbe6..f2ce8da7 100644 --- a/src/observer/sql/executor/load_data_executor.cpp +++ b/src/observer/sql/executor/load_data_executor.cpp @@ -26,7 +26,7 @@ RC LoadDataExecutor::execute(SQLStageEvent *sql_event) RC rc = RC::SUCCESS; SqlResult *sql_result = sql_event->session_event()->sql_result(); LoadDataStmt *stmt = static_cast(sql_event->stmt()); - Table *table = stmt->table(); + auto table = stmt->table(); const char *file_name = stmt->filename(); load_data(table, file_name, sql_result); return rc; @@ -40,8 +40,8 @@ RC LoadDataExecutor::execute(SQLStageEvent *sql_event) * @param errmsg 如果出现错误，通过这个参数返回错误信息 * @return 成功返回RC::SUCCESS */ -RC insert_record_from_file( - Table *table, std::vector &file_values, std::vector &record_values, std::stringstream &errmsg) +RC insert_record_from_file(BaseTable *table, std::vector &file_values, std::vector &record_values, + std::stringstream &errmsg) { const int field_num = record_values.size(); @@ -76,7 +76,7 @@ RC insert_record_from_file( return rc; } -void LoadDataExecutor::load_data(Table *table, const char *file_name, SqlResult *sql_result) +void LoadDataExecutor::load_data(BaseTable *table, const char *file_name, SqlResult *sql_result) { std::stringstream result_string; diff --git a/src/observer/sql/executor/load_data_executor.h b/src/observer/sql/executor/load_data_executor.h index dc023ba9..8fce1641 100644 --- a/src/observer/sql/executor/load_data_executor.h +++ b/src/observer/sql/executor/load_data_executor.h @@ -16,8 +16,8 @@ See the Mulan PSL v2 for more details. */ #include "common/rc.h" +class BaseTable; class SQLStageEvent; -class Table; class SqlResult; /** @@ -33,5 +33,5 @@ class LoadDataExecutor RC execute(SQLStageEvent *sql_event); private: - void load_data(Table *table, const char *file_name, SqlResult *sql_result); + void load_data(BaseTable *table, const char *file_name, SqlResult *sql_result); }; diff --git a/src/observer/sql/executor/set_variable_executor.cpp b/src/observer/sql/executor/set_variable_executor.cpp index b5518a18..2279f12b 100644 --- a/src/observer/sql/executor/set_variable_executor.cpp +++ b/src/observer/sql/executor/set_variable_executor.cpp @@ -12,87 +12,87 @@ See the Mulan PSL v2 for more details. */ RC SetVariableExecutor::execute(SQLStageEvent *sql_event) { - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; - Session *session = sql_event->session_event()->session(); - SetVariableStmt *stmt = (SetVariableStmt *)sql_event->stmt(); + Session *session = sql_event->session_event()->session(); + SetVariableStmt *stmt = (SetVariableStmt *)sql_event->stmt(); - const char *var_name = stmt->var_name(); - const Value &var_value = stmt->var_value(); - if (strcasecmp(var_name, "sql_debug") == 0) { - bool bool_value = false; - rc = var_value_to_boolean(var_value, bool_value); - if (rc == RC::SUCCESS) { - session->set_sql_debug(bool_value); - LOG_TRACE("set sql_debug to %d", bool_value); - } - } else if (strcasecmp(var_name, "execution_mode") == 0) { - ExecutionMode execution_mode = ExecutionMode::UNKNOWN_MODE; - rc = get_execution_mode(var_value, execution_mode); - if (rc == RC::SUCCESS && execution_mode != ExecutionMode::UNKNOWN_MODE) { - session->set_execution_mode(execution_mode); - } else { - rc = RC::INVALID_ARGUMENT; - } + const char *var_name = stmt->var_name(); + const Value &var_value = stmt->var_value(); + if (strcasecmp(var_name, "sql_debug") == 0) { + bool bool_value = false; + rc = var_value_to_boolean(var_value, bool_value); + if (rc == RC::SUCCESS) { + session->set_sql_debug(bool_value); + LOG_TRACE("set sql_debug to %d", bool_value); + } + } else if (strcasecmp(var_name, "execution_mode") == 0) { + ExecutionMode execution_mode = ExecutionMode::UNKNOWN_MODE; + rc = get_execution_mode(var_value, execution_mode); + if (rc == RC::SUCCESS && execution_mode != ExecutionMode::UNKNOWN_MODE) { + session->set_execution_mode(execution_mode); } else { - rc = RC::VARIABLE_NOT_EXISTS; + rc = RC::INVALID_ARGUMENT; } + } else { + rc = RC::VARIABLE_NOT_EXISTS; + } - return rc; + return rc; } RC SetVariableExecutor::var_value_to_boolean(const Value &var_value, bool &bool_value) const { - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; - if (var_value.attr_type() == AttrType::BOOLEANS) { - bool_value = var_value.get_boolean(); - } else if (var_value.attr_type() == AttrType::INTS) { - bool_value = var_value.get_int() != 0; - } else if (var_value.attr_type() == AttrType::FLOATS) { - bool_value = var_value.get_float() != 0.0; - } else if (var_value.attr_type() == AttrType::CHARS) { + if (var_value.attr_type() == AttrType::BOOLEANS) { + bool_value = var_value.get_boolean(); + } else if (var_value.attr_type() == AttrType::INTS) { + bool_value = var_value.get_int() != 0; + } else if (var_value.attr_type() == AttrType::FLOATS) { + bool_value = var_value.get_float() != 0.0; + } else if (var_value.attr_type() == AttrType::CHARS) { - std::string true_strings[] = {"true", "on", "yes", "t", "1"}; + std::string true_strings[] = {"true", "on", "yes", "t", "1"}; - std::string false_strings[] = {"false", "off", "no", "f", "0"}; + std::string false_strings[] = {"false", "off", "no", "f", "0"}; - for (size_t i = 0; i < sizeof(true_strings) / sizeof(true_strings[0]); i++) { - if (strcasecmp(var_value.get_string().c_str(), true_strings[i].c_str()) == 0) { - bool_value = true; - return rc; - } + for (size_t i = 0; i < sizeof(true_strings) / sizeof(true_strings[0]); i++) { + if (strcasecmp(var_value.get_string().c_str(), true_strings[i].c_str()) == 0) { + bool_value = true; + return rc; } + } - for (size_t i = 0; i < sizeof(false_strings) / sizeof(false_strings[0]); i++) { - if (strcasecmp(var_value.get_string().c_str(), false_strings[i].c_str()) == 0) { - bool_value = false; - return rc; - } + for (size_t i = 0; i < sizeof(false_strings) / sizeof(false_strings[0]); i++) { + if (strcasecmp(var_value.get_string().c_str(), false_strings[i].c_str()) == 0) { + bool_value = false; + return rc; } - rc = RC::VARIABLE_NOT_VALID; } + rc = RC::VARIABLE_NOT_VALID; + } - return rc; + return rc; } RC SetVariableExecutor::get_execution_mode(const Value &var_value, ExecutionMode &execution_mode) const { - RC rc = RC::SUCCESS; - - if (var_value.attr_type() == AttrType::CHARS) { - if (strcasecmp(var_value.get_string().c_str(), "TUPLE_ITERATOR") == 0) { - execution_mode = ExecutionMode::TUPLE_ITERATOR; - } else if (strcasecmp(var_value.get_string().c_str(), "CHUNK_ITERATOR") == 0) { - execution_mode = ExecutionMode::CHUNK_ITERATOR; - } else { - execution_mode = ExecutionMode::UNKNOWN_MODE; - rc = RC::VARIABLE_NOT_VALID; - } - } else { + RC rc = RC::SUCCESS; + + if (var_value.attr_type() == AttrType::CHARS) { + if (strcasecmp(var_value.get_string().c_str(), "TUPLE_ITERATOR") == 0) { + execution_mode = ExecutionMode::TUPLE_ITERATOR; + } else if (strcasecmp(var_value.get_string().c_str(), "CHUNK_ITERATOR") == 0) { + execution_mode = ExecutionMode::CHUNK_ITERATOR; + } else { execution_mode = ExecutionMode::UNKNOWN_MODE; - rc = RC::VARIABLE_NOT_VALID; + rc = RC::VARIABLE_NOT_VALID; } + } else { + execution_mode = ExecutionMode::UNKNOWN_MODE; + rc = RC::VARIABLE_NOT_VALID; + } - return rc; + return rc; } \ No newline at end of file diff --git a/src/observer/sql/executor/show_index_executor.cpp b/src/observer/sql/executor/show_index_executor.cpp new file mode 100644 index 00000000..9762911e --- /dev/null +++ b/src/observer/sql/executor/show_index_executor.cpp @@ -0,0 +1,70 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/29 * + * @Description : ShowIndexExecutor source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "sql/executor/show_index_executor.h" +#include "sql/stmt/stmt.h" +#include "event/session_event.h" +#include "common/log/log.h" +#include "event/sql_event.h" +#include "session/session.h" +#include "sql/operator/string_list_physical_operator.h" +#include "sql/stmt/desc_table_stmt.h" +#include "storage/db/db.h" +#include "storage/table/table.h" +#include "sql/stmt/show_index_stmt.h" + +RC ShowIndexExecutor::execute(SQLStageEvent *sql_event) +{ + RC rc = RC::SUCCESS; + Stmt *stmt = sql_event->stmt(); + SessionEvent *session_event = sql_event->session_event(); + Session *session = session_event->session(); + ASSERT(stmt->type() == StmtType::SHOW_INDEX, + "show index executor can not run this command: %d", + static_cast(stmt->type())); + + ShowIndexStmt *show_index_stmt = static_cast(stmt); + SqlResult *sql_result = session_event->sql_result(); + const char *table_name = show_index_stmt->table_name().c_str(); + + Db *db = session->get_current_db(); + BaseTable *table = db->find_table(table_name); + if (table != nullptr) { + TupleSchema tuple_schema; + tuple_schema.append_cell(TupleCellSpec("", "Table", "Table")); + tuple_schema.append_cell(TupleCellSpec("", "Non_unique", "Non_unique")); + tuple_schema.append_cell(TupleCellSpec("", "Key_name", "Key_name")); + tuple_schema.append_cell(TupleCellSpec("", "Seq_in_index", "Seq_in_index")); + tuple_schema.append_cell(TupleCellSpec("", "Column_name", "Column_name")); + sql_result->set_tuple_schema(tuple_schema); + + auto oper = new StringListPhysicalOperator; + const TableMeta &table_meta = table->table_meta(); + for (int i = 0; i < table_meta.index_num(); i++) { + auto index = table_meta.index(i); + for (size_t j = 0; j < index->fields().size(); j++) { + auto unique = index->unique() ? "0" : "1"; + auto name = index->name(); + auto id = std::to_string(j + 1); + auto field_name = index->fields()[j].name(); + vector list = {table_name, unique, name, id, field_name}; + oper->append(list); + } + } + + sql_result->set_operator(unique_ptr(oper)); + } else { + sql_result->set_return_code(RC::SCHEMA_TABLE_NOT_EXIST); + sql_result->set_state_string("Table not exists"); + } + return rc; +} diff --git a/src/observer/sql/executor/show_index_executor.h b/src/observer/sql/executor/show_index_executor.h new file mode 100644 index 00000000..75331b8c --- /dev/null +++ b/src/observer/sql/executor/show_index_executor.h @@ -0,0 +1,30 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/29 * + * @Description : ShowIndexExecutor header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "common/rc.h" + +class SQLStageEvent; + +/** + * @brief 描述表的执行器 + * @ingroup Executor + */ +class ShowIndexExecutor +{ +public: + ShowIndexExecutor() = default; + virtual ~ShowIndexExecutor() = default; + + RC execute(SQLStageEvent *sql_event); +}; diff --git a/src/observer/sql/executor/sql_result.cpp b/src/observer/sql/executor/sql_result.cpp index a8b1b855..11ca7eb6 100644 --- a/src/observer/sql/executor/sql_result.cpp +++ b/src/observer/sql/executor/sql_result.cpp @@ -79,5 +79,6 @@ void SqlResult::set_operator(std::unique_ptr oper) { ASSERT(operator_ == nullptr, "current operator is not null. Result is not closed?"); operator_ = std::move(oper); + // 这里调用了投影算子的方法，补充表头 operator_->tuple_schema(tuple_schema_); } diff --git a/src/observer/sql/expr/aggregate_hash_table.cpp b/src/observer/sql/expr/aggregate_hash_table.cpp index cde07fd4..1367c646 100644 --- a/src/observer/sql/expr/aggregate_hash_table.cpp +++ b/src/observer/sql/expr/aggregate_hash_table.cpp @@ -169,7 +169,7 @@ RC LinearProbingAggregateHashTable::iter_get(int pos, int &key, V &value) template void LinearProbingAggregateHashTable::aggregate(V *value, V value_to_aggregate) { - if (aggregate_type_ == AggregateExpr::Type::SUM) { + if (aggregate_type_ == AggregateFunctionExpr::Type::SUM) { *value += value_to_aggregate; } else { ASSERT(false, "unsupported aggregate type"); @@ -222,16 +222,18 @@ void LinearProbingAggregateHashTable::add_batch(int *input_keys, V *input_val // off 全部初始化为 0 // for (; i + SIMD_WIDTH <= len;) { - // 1: 根据 `inv` 变量的值，从 `input_keys` 中 `selective load` `SIMD_WIDTH` 个不同的输入键值对。 - // 2. 计算 i += |inv|, `|inv|` 表示 `inv` 中有效的个数 - // 3. 计算 hash 值， - // 4. 根据聚合类型（目前只支持 sum），在哈希表中更新聚合结果。如果本次循环，没有找到key[i] 在哈希表中的位置，则不更新聚合结果。 - // 5. gather 操作，根据 hash 值将 keys_ 的 gather 结果写入 table_key 中。 - // 6. 更新 inv 和 off。如果本次循环key[i] 聚合完成，则inv[i]=-1，表示该位置在下次循环中读取新的键值对。 - // 如果本次循环 key[i] 未在哈希表中聚合完成（table_key[i] != key[i]），则inv[i] = 0，表示该位置在下次循环中不需要读取新的键值对。 - // 如果本次循环中，key[i]聚合完成，则off[i] 更新为 0，表示线性探测偏移量为 0，key[i] 未完成聚合，则off[i]++,表示线性探测偏移量加 1。 + // 1: 根据 `inv` 变量的值，从 `input_keys` 中 `selective load` `SIMD_WIDTH` 个不同的输入键值对。 + // 2. 计算 i += |inv|, `|inv|` 表示 `inv` 中有效的个数 + // 3. 计算 hash 值， + // 4. 根据聚合类型（目前只支持 sum），在哈希表中更新聚合结果。如果本次循环，没有找到key[i] + // 在哈希表中的位置，则不更新聚合结果。 + // 5. gather 操作，根据 hash 值将 keys_ 的 gather 结果写入 table_key 中。 + // 6. 更新 inv 和 off。如果本次循环key[i] 聚合完成，则inv[i]=-1，表示该位置在下次循环中读取新的键值对。 + // 如果本次循环 key[i] 未在哈希表中聚合完成（table_key[i] != key[i]），则inv[i] = + // 0，表示该位置在下次循环中不需要读取新的键值对。 如果本次循环中，key[i]聚合完成，则off[i] 更新为 + // 0，表示线性探测偏移量为 0，key[i] 未完成聚合，则off[i]++,表示线性探测偏移量加 1。 // } - //7. 通过标量线性探测，处理剩余键值对 + // 7. 通过标量线性探测，处理剩余键值对 // resize_if_need(); } diff --git a/src/observer/sql/expr/aggregate_hash_table.h b/src/observer/sql/expr/aggregate_hash_table.h index 89719c9b..1a64a299 100644 --- a/src/observer/sql/expr/aggregate_hash_table.h +++ b/src/observer/sql/expr/aggregate_hash_table.h @@ -35,7 +35,7 @@ class AggregateHashTable */ virtual RC next(Chunk &chunk) = 0; - virtual void close_scan(){}; + virtual void close_scan() {}; protected: AggregateHashTable *hash_table_; @@ -82,7 +82,7 @@ class StandardAggregateHashTable : public AggregateHashTable { for (auto &expr : aggregations) { ASSERT(expr->type() == ExprType::AGGREGATION, "expect aggregate expression"); - auto *aggregation_expr = static_cast(expr); + auto *aggregation_expr = static_cast(expr); aggr_types_.push_back(aggregation_expr->aggregate_type()); } } @@ -96,8 +96,8 @@ class StandardAggregateHashTable : public AggregateHashTable private: /// group by values -> aggregate values - StandardHashTable aggr_values_; - std::vector aggr_types_; + StandardHashTable aggr_values_; + std::vector aggr_types_; }; /** @@ -128,7 +128,7 @@ class LinearProbingAggregateHashTable : public AggregateHashTable int scan_count_ = 0; }; - LinearProbingAggregateHashTable(AggregateExpr::Type aggregate_type, int capacity = DEFAULT_CAPACITY) + LinearProbingAggregateHashTable(AggregateFunctionExpr::Type aggregate_type, int capacity = DEFAULT_CAPACITY) : keys_(capacity, EMPTY_KEY), values_(capacity, 0), capacity_(capacity), aggregate_type_(aggregate_type) {} virtual ~LinearProbingAggregateHashTable() {} @@ -162,10 +162,10 @@ class LinearProbingAggregateHashTable : public AggregateHashTable static const int EMPTY_KEY; static const int DEFAULT_CAPACITY; - std::vector keys_; - std::vector values_; - int size_ = 0; - int capacity_ = 0; - AggregateExpr::Type aggregate_type_; + std::vector keys_; + std::vector values_; + int size_ = 0; + int capacity_ = 0; + AggregateFunctionExpr::Type aggregate_type_; }; #endif \ No newline at end of file diff --git a/src/observer/sql/expr/aggregate_state.cpp b/src/observer/sql/expr/aggregate_state.cpp index 124cb277..0e394ce8 100644 --- a/src/observer/sql/expr/aggregate_state.cpp +++ b/src/observer/sql/expr/aggregate_state.cpp @@ -26,7 +26,7 @@ void SumState::update(const T *values, int size) } #else for (int i = 0; i < size; ++i) { - value += values[i]; + value += values[i]; } #endif } diff --git a/src/observer/sql/expr/aggregator.h b/src/observer/sql/expr/aggregator.h index c1af3d65..4f1b8d9d 100644 --- a/src/observer/sql/expr/aggregator.h +++ b/src/observer/sql/expr/aggregator.h @@ -2,16 +2,12 @@ miniob is licensed under Mulan PSL v2. You can use this software according to the terms and conditions of the Mulan PSL v2. You may obtain a copy of Mulan PSL v2 at: - http://license.coscl.org.cn/MulanPSL2 + http: THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. See the Mulan PSL v2 for more details. */ -// -// Created by Wangyunlai on 2024/05/29. -// - #pragma once #include "common/value.h" @@ -26,12 +22,96 @@ class Aggregator virtual RC evaluate(Value &result) = 0; protected: - Value value_; + Value value_ = Value(NullValue()); }; -class SumAggregator : public Aggregator +class CountAggregator : public Aggregator { public: - RC accumulate(const Value &value) override; - RC evaluate(Value &result) override; -}; \ No newline at end of file + RC accumulate(const Value &value) override + { + if (value.is_null()) { + return RC::SUCCESS; + } + count_++; + return RC::SUCCESS; + } + + RC evaluate(Value &result) override + { + result = Value(count_); + return RC::SUCCESS; + } + +private: + int count_ = 0; +}; + +class AvgAggregator : public Aggregator +{ +public: + RC accumulate(const Value &value) override + { + if (value.is_null()) { + return RC::SUCCESS; + } + if (value_.is_null()) { + value_ = value; + count_ = 1; + } else { + Value::add(value, value_, value_); + count_++; + } + return RC::SUCCESS; + } + + RC evaluate(Value &result) override + { + if (count_ > 0) { + Value avg = value_; + avg = Value(avg.get_float() / static_cast(count_)); + result = avg; + } else { + result = Value(NullValue()); + } + return RC::SUCCESS; + } + +private: + int count_ = 0; +}; + +#define _agg(FUN) \ +public: \ + RC accumulate(const Value &value) override \ + { \ + if (value.is_null()) { \ + return RC::SUCCESS; \ + } \ + if (value_.is_null()) { \ + value_ = value; \ + } else { \ + FUN; \ + } \ + return RC::SUCCESS; \ + } \ + RC evaluate(Value &result) override \ + { \ + result = value_; \ + return RC::SUCCESS; \ + } + +class SumAggregator : public Aggregator +{ + _agg(Value::add(value, value_, value_)) +}; + +class MaxAggregator : public Aggregator +{ + _agg(if (value.compare(value_) > 0) { value_ = value; }) +}; + +class MinAggregator : public Aggregator +{ + _agg(if (value.compare(value_) < 0) { value_ = value; }) +}; diff --git a/src/observer/sql/expr/arithmetic_operator.hpp b/src/observer/sql/expr/arithmetic_operator.hpp index 2b15e2ef..fa625f76 100644 --- a/src/observer/sql/expr/arithmetic_operator.hpp +++ b/src/observer/sql/expr/arithmetic_operator.hpp @@ -208,7 +208,7 @@ template void compare_operation(T *left, T *right, int n, std::vector &result) { #if defined(USE_SIMD) - int i = 0; + int i = 0; if constexpr (std::is_same::value) { for (; i <= n - SIMD_WIDTH; i += SIMD_WIDTH) { __m256 left_value, right_value; diff --git a/src/observer/sql/expr/composite_tuple.cpp b/src/observer/sql/expr/composite_tuple.cpp index 40ff8128..94c82857 100644 --- a/src/observer/sql/expr/composite_tuple.cpp +++ b/src/observer/sql/expr/composite_tuple.cpp @@ -64,8 +64,17 @@ RC CompositeTuple::find_cell(const TupleCellSpec &spec, Value &cell) const void CompositeTuple::add_tuple(unique_ptr tuple) { tuples_.push_back(std::move(tuple)); } -Tuple &CompositeTuple::tuple_at(size_t index) -{ +Tuple &CompositeTuple::tuple_at(size_t index) +{ ASSERT(index < tuples_.size(), "index=%d, tuples_size=%d", index, tuples_.size()); - return *tuples_[index]; + return *tuples_[index]; +} + +Tuple *CompositeTuple::copy() const +{ + CompositeTuple *copy = new CompositeTuple(); + for (auto &tuple : tuples_) { + copy->tuples_.emplace_back(tuple->copy()); + } + return copy; } \ No newline at end of file diff --git a/src/observer/sql/expr/composite_tuple.h b/src/observer/sql/expr/composite_tuple.h index d3e65577..23864bbe 100644 --- a/src/observer/sql/expr/composite_tuple.h +++ b/src/observer/sql/expr/composite_tuple.h @@ -45,6 +45,8 @@ class CompositeTuple : public Tuple void add_tuple(std::unique_ptr tuple); Tuple &tuple_at(size_t index); + Tuple *copy() const override; + private: std::vector> tuples_; }; \ No newline at end of file diff --git a/src/observer/sql/expr/expression.cpp b/src/observer/sql/expr/expression.cpp index 1c1aa48c..2ad28a46 100644 --- a/src/observer/sql/expr/expression.cpp +++ b/src/observer/sql/expr/expression.cpp @@ -16,11 +16,22 @@ See the Mulan PSL v2 for more details. */ #include "sql/expr/tuple.h" #include "sql/expr/arithmetic_operator.hpp" -using namespace std; +#include "sql/stmt/select_stmt.h" +#include "sql/operator/logical_operator.h" +#include "sql/operator/physical_operator.h" +#include "sql/optimizer/logical_plan_generator.h" +#include "sql/optimizer/physical_plan_generator.h" +#include "common/lang/defer.h" -RC FieldExpr::get_value(const Tuple &tuple, Value &value) const +#define check_type(str, rule) \ + if (0 == strcasecmp(type_str, str)) { \ + type = rule; \ + return RC::SUCCESS; \ + } + +RC FieldExpr::get_value(const Tuple &tuple, Value &value) { - return tuple.find_cell(TupleCellSpec(table_name(), field_name()), value); + return tuple.find_cell(TupleCellSpec(table_name(), field_name(), table_alias_.c_str()), value); } bool FieldExpr::equal(const Expression &other) const @@ -59,7 +70,7 @@ bool ValueExpr::equal(const Expression &other) const return value_.compare(other_value_expr.get_value()) == 0; } -RC ValueExpr::get_value(const Tuple &tuple, Value &value) const +RC ValueExpr::get_value(const Tuple &tuple, Value &value) { value = value_; return RC::SUCCESS; @@ -88,10 +99,10 @@ RC CastExpr::cast(const Value &value, Value &cast_value) const return rc; } -RC CastExpr::get_value(const Tuple &tuple, Value &result) const +RC CastExpr::get_value(const Tuple &tuple, Value &result) { Value value; - RC rc = child_->get_value(tuple, value); + RC rc = child_->get_value(tuple, value); if (rc != RC::SUCCESS) { return rc; } @@ -102,7 +113,7 @@ RC CastExpr::get_value(const Tuple &tuple, Value &result) const RC CastExpr::try_get_value(Value &result) const { Value value; - RC rc = child_->try_get_value(value); + RC rc = child_->try_get_value(value); if (rc != RC::SUCCESS) { return rc; } @@ -112,15 +123,45 @@ RC CastExpr::try_get_value(Value &result) const //////////////////////////////////////////////////////////////////////////////// -ComparisonExpr::ComparisonExpr(CompOp comp, unique_ptr left, unique_ptr right) +ComparisonExpr::ComparisonExpr(CompOp comp, Expression *left, Expression *right) + : comp_(comp), left_(std::unique_ptr(left)), right_(std::unique_ptr(right)) +{} + +ComparisonExpr::ComparisonExpr(CompOp comp, std::unique_ptr left, std::unique_ptr right) : comp_(comp), left_(std::move(left)), right_(std::move(right)) {} -ComparisonExpr::~ComparisonExpr() {} +ComparisonExpr::~ComparisonExpr() = default; RC ComparisonExpr::compare_value(const Value &left, const Value &right, bool &result) const { - RC rc = RC::SUCCESS; + if (comp_ == IS_OP) { + if (right.attr_type() != AttrType::NULLS) { + return RC::NOT_NULL_AFTER_IS; + } + result = left.is_null(); + return RC::SUCCESS; + } + if (comp_ == IS_NOT_OP) { + if (right.attr_type() != AttrType::NULLS) { + return RC::NOT_NULL_AFTER_IS; + } + result = !left.is_null(); + return RC::SUCCESS; + } + if (left.is_null() || right.is_null()) { + result = false; + return RC::SUCCESS; + } + + RC rc = RC::SUCCESS; + + if (comp_ == LIKE_OP || comp_ == NOT_LIKE_OP) { + ASSERT(left.is_str() && right.is_str(), "LIKE ONLY SUPPORT STRING TYPE!"); + result = comp_ == LIKE_OP ? left.LIKE(right) : !left.LIKE(right); + return rc; + } + int cmp_result = left.compare(right); result = false; switch (comp_) { @@ -154,8 +195,8 @@ RC ComparisonExpr::compare_value(const Value &left, const Value &right, bool &re RC ComparisonExpr::try_get_value(Value &cell) const { if (left_->type() == ExprType::VALUE && right_->type() == ExprType::VALUE) { - ValueExpr * left_value_expr = static_cast(left_.get()); - ValueExpr * right_value_expr = static_cast(right_.get()); + ValueExpr *left_value_expr = static_cast(left_.get()); + ValueExpr *right_value_expr = static_cast(right_.get()); const Value &left_cell = left_value_expr->get_value(); const Value &right_cell = right_value_expr->get_value(); @@ -172,25 +213,111 @@ RC ComparisonExpr::try_get_value(Value &cell) const return RC::INVALID_ARGUMENT; } -RC ComparisonExpr::get_value(const Tuple &tuple, Value &value) const +RC ComparisonExpr::get_value(const Tuple &tuple, Value &value) { + RC rc = RC::SUCCESS; Value left_value; Value right_value; - RC rc = left_->get_value(tuple, left_value); - if (rc != RC::SUCCESS) { + SubQueryExpr *left_subquery_expr = nullptr; + SubQueryExpr *right_subquery_expr = nullptr; + + // Lambda to check if the expression is a subquery and open it + auto open_subquery = [&tuple](const std::unique_ptr &expr, SubQueryExpr *&subquery_expr) -> RC { + if (expr->type() == ExprType::SUBQUERY) { + subquery_expr = dynamic_cast(expr.get()); + RC rc = subquery_expr->open(nullptr, tuple); // Open the subquery expression (pass nullptr for now) + if (OB_FAIL(rc)) { + return rc; + } + } + return RC::SUCCESS; + }; + + // Check and open the left subquery expression if it exists + rc = open_subquery(left_, left_subquery_expr); + if (OB_FAIL(rc)) { + LOG_WARN("failed to open left subquery expression. rc=%s", strrc(rc)); + return rc; + } + + // Check and open the right subquery expression if it exists + rc = open_subquery(right_, right_subquery_expr); + if (OB_FAIL(rc)) { + LOG_WARN("failed to open right subquery expression. rc=%s", strrc(rc)); + return rc; + } + DEFER(if (nullptr != left_subquery_expr) left_subquery_expr->close(); + if (nullptr != right_subquery_expr) right_subquery_expr->close();); + + // Get the value of the left expression + rc = left_->get_value(tuple, left_value); + if (rc != RC::SUCCESS && rc != RC::RECORD_EOF) { LOG_WARN("failed to get value of left expression. rc=%s", strrc(rc)); return rc; } + + // Check if the left subquery has more rows (error if true) + if (left_subquery_expr && left_subquery_expr->has_more_row(tuple)) { + return RC::SUBQUERY_RETURNED_MULTIPLE_ROWS; + } + + // Handle IN and NOT IN operations + if (comp_ == IN_OP || comp_ == NOT_IN_OP) { + if (left_value.is_null()) { + value.set_boolean(false); + return RC::SUCCESS; + } + + if (right_->type() == ExprType::EXPRLIST) { + static_cast(right_.get())->reset(); + } + + // 比较表达式的结果，如果进入 while 循环且没有提前退出，那么结果即为该值 + bool res = comp_ == NOT_IN_OP; + + rc = right_->get_value(tuple, right_value); + if (rc == RC::RECORD_EOF) { + // 子查询结果为空，返回 null 值 + } else if (OB_FAIL(rc)) { + // 其他错误 + return rc; + } else if (left_value.compare(right_value) == 0) { + // 不为空才能比较，null 是不可比较的 + res = comp_ == IN_OP; + } else { + while (RC::SUCCESS == (rc = right_->get_value(tuple, right_value))) { + if (right_value.is_null()) { + // 对于 not in，一边有 null 就为假 + if (comp_ == NOT_IN_OP) { + res = false; + break; + } + } else if (left_value.compare(right_value) == 0) { + res = comp_ == IN_OP; + break; + } + } + } + value.set_boolean(res); + return rc == RC::RECORD_EOF ? RC::SUCCESS : rc; + } + + // Get the value of the right expression rc = right_->get_value(tuple, right_value); - if (rc != RC::SUCCESS) { + if (rc != RC::SUCCESS && rc != RC::RECORD_EOF) { LOG_WARN("failed to get value of right expression. rc=%s", strrc(rc)); return rc; } - bool bool_value = false; + // Check if the right subquery has more rows (error if true) + if (right_subquery_expr && right_subquery_expr->has_more_row(tuple)) { + return RC::SUBQUERY_RETURNED_MULTIPLE_ROWS; + } - rc = compare_value(left_value, right_value, bool_value); + // Compare the left and right values + bool bool_value = false; + rc = compare_value(left_value, right_value, bool_value); if (rc == RC::SUCCESS) { value.set_boolean(bool_value); } @@ -253,7 +380,18 @@ ConjunctionExpr::ConjunctionExpr(Type type, vector> &chil : conjunction_type_(type), children_(std::move(children)) {} -RC ConjunctionExpr::get_value(const Tuple &tuple, Value &value) const +ConjunctionExpr::ConjunctionExpr(Type type, Expression *left, Expression *right) : conjunction_type_(type) +{ + children_.emplace_back(left); + children_.emplace_back(right); +} + +ConjunctionExpr::ConjunctionExpr(Type type, std::unique_ptr children) : conjunction_type_(type) +{ + children_.push_back(std::move(children)); +} + +RC ConjunctionExpr::get_value(const Tuple &tuple, Value &value) { RC rc = RC::SUCCESS; if (children_.empty()) { @@ -312,40 +450,47 @@ AttrType ArithmeticExpr::value_type() const return AttrType::INTS; } + if (left_->value_type() == AttrType::VECTORS && right_->value_type() == AttrType::VECTORS) { + return AttrType::VECTORS; + } + return AttrType::FLOATS; } RC ArithmeticExpr::calc_value(const Value &left_value, const Value &right_value, Value &value) const { RC rc = RC::SUCCESS; + if (left_value.is_null() || right_value.is_null()) { + value.set_null(true); + return RC::SUCCESS; + } const AttrType target_type = value_type(); value.set_type(target_type); switch (arithmetic_type_) { case Type::ADD: { - Value::add(left_value, right_value, value); + return Value::add(left_value, right_value, value); } break; case Type::SUB: { - Value::subtract(left_value, right_value, value); + return Value::subtract(left_value, right_value, value); } break; case Type::MUL: { - Value::multiply(left_value, right_value, value); + return Value::multiply(left_value, right_value, value); } break; case Type::DIV: { - Value::divide(left_value, right_value, value); + return Value::divide(left_value, right_value, value); } break; case Type::NEGATIVE: { - Value::negative(left_value, value); + return Value::negative(left_value, value); } break; default: { - rc = RC::INTERNAL; - LOG_WARN("unsupported arithmetic type. %d", arithmetic_type_); + return RC::INTERNAL; } break; } return rc; @@ -419,22 +564,26 @@ RC ArithmeticExpr::execute_calc( return rc; } -RC ArithmeticExpr::get_value(const Tuple &tuple, Value &value) const +RC ArithmeticExpr::get_value(const Tuple &tuple, Value &value) { RC rc = RC::SUCCESS; Value left_value; Value right_value; - rc = left_->get_value(tuple, left_value); - if (rc != RC::SUCCESS) { - LOG_WARN("failed to get value of left expression. rc=%s", strrc(rc)); - return rc; + if (left_) { + rc = left_->get_value(tuple, left_value); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to get value of left expression. rc=%s", strrc(rc)); + return rc; + } } - rc = right_->get_value(tuple, right_value); - if (rc != RC::SUCCESS) { - LOG_WARN("failed to get value of right expression. rc=%s", strrc(rc)); - return rc; + if (right_) { + rc = right_->get_value(tuple, right_value); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to get value of right expression. rc=%s", strrc(rc)); + return rc; + } } return calc_value(left_value, right_value, value); } @@ -512,17 +661,20 @@ RC ArithmeticExpr::try_get_value(Value &value) const //////////////////////////////////////////////////////////////////////////////// -UnboundAggregateExpr::UnboundAggregateExpr(const char *aggregate_name, Expression *child) - : aggregate_name_(aggregate_name), child_(child) +UnboundFunctionExpr::UnboundFunctionExpr(const char *aggregate_name, std::vector> child) + : function_name_(aggregate_name), args_(std::move(child)) {} //////////////////////////////////////////////////////////////////////////////// -AggregateExpr::AggregateExpr(Type type, Expression *child) : aggregate_type_(type), child_(child) {} +AggregateFunctionExpr::AggregateFunctionExpr(AggregateFunctionType type, Expression *child) + : aggregate_type_(type), child_(child) +{} -AggregateExpr::AggregateExpr(Type type, unique_ptr child) : aggregate_type_(type), child_(std::move(child)) +AggregateFunctionExpr::AggregateFunctionExpr(AggregateFunctionType type, unique_ptr child) + : aggregate_type_(type), child_(std::move(child)) {} -RC AggregateExpr::get_column(Chunk &chunk, Column &column) +RC AggregateFunctionExpr::get_column(Chunk &chunk, Column &column) { RC rc = RC::SUCCESS; if (pos_ != -1) { @@ -533,7 +685,7 @@ RC AggregateExpr::get_column(Chunk &chunk, Column &column) return rc; } -bool AggregateExpr::equal(const Expression &other) const +bool AggregateFunctionExpr::equal(const Expression &other) const { if (this == &other) { return true; @@ -541,18 +693,34 @@ bool AggregateExpr::equal(const Expression &other) const if (other.type() != type()) { return false; } - const AggregateExpr &other_aggr_expr = static_cast(other); + const AggregateFunctionExpr &other_aggr_expr = static_cast(other); return aggregate_type_ == other_aggr_expr.aggregate_type() && child_->equal(*other_aggr_expr.child()); } -unique_ptr AggregateExpr::create_aggregator() const +unique_ptr AggregateFunctionExpr::create_aggregator() const { unique_ptr aggregator; switch (aggregate_type_) { - case Type::SUM: { + case AggregateFunctionType::SUM: { aggregator = make_unique(); break; } + case AggregateFunctionType::COUNT: { + aggregator = make_unique(); + break; + } + case AggregateFunctionType::AVG: { + aggregator = make_unique(); + break; + } + case AggregateFunctionType::MAX: { + aggregator = make_unique(); + break; + } + case AggregateFunctionType::MIN: { + aggregator = make_unique(); + break; + } default: { ASSERT(false, "unsupported aggregate type"); break; @@ -561,26 +729,234 @@ unique_ptr AggregateExpr::create_aggregator() const return aggregator; } -RC AggregateExpr::get_value(const Tuple &tuple, Value &value) const +RC AggregateFunctionExpr::get_value(const Tuple &tuple, Value &value) { return tuple.find_cell(TupleCellSpec(name()), value); } -RC AggregateExpr::type_from_string(const char *type_str, AggregateExpr::Type &type) +RC AggregateFunctionExpr::type_from_string(const char *type_str, AggregateFunctionType &type) { - RC rc = RC::SUCCESS; - if (0 == strcasecmp(type_str, "count")) { - type = Type::COUNT; - } else if (0 == strcasecmp(type_str, "sum")) { - type = Type::SUM; - } else if (0 == strcasecmp(type_str, "avg")) { - type = Type::AVG; - } else if (0 == strcasecmp(type_str, "max")) { - type = Type::MAX; - } else if (0 == strcasecmp(type_str, "min")) { - type = Type::MIN; - } else { - rc = RC::INVALID_ARGUMENT; + check_type("sum", AggregateFunctionType::SUM); + check_type("avg", AggregateFunctionType::AVG); + check_type("max", AggregateFunctionType::MAX); + check_type("min", AggregateFunctionType::MIN); + check_type("count", AggregateFunctionType::COUNT); + return RC::INVALID_ARGUMENT; +} + +SubQueryExpr::SubQueryExpr(SelectSqlNode &select_node) : sql_node_(select_node) {} + +SubQueryExpr::~SubQueryExpr() = default; + +RC SubQueryExpr::generate_select_stmt(Db *db, const std::unordered_map &tables) +{ + // 仿照普通 select 的执行流程，tables 用来传递别名 + Stmt *stmt = nullptr; + RC rc = SelectStmt::create(db, sql_node_, stmt, tables); + if (OB_FAIL(rc)) { + LOG_WARN("failed to create subquery select statement. return %s", strrc(rc)); + return rc; + } + + // 确保生成的 stmt 类型为 SELECT 类型 + if (stmt->type() != StmtType::SELECT) { + LOG_WARN("subquery stmt type is not SELECT."); + return RC::INVALID_ARGUMENT; + } + + // 动态转换为 SelectStmt 类型，并进行子查询列数校验 + auto *select_stmt = dynamic_cast(stmt); + if (select_stmt == nullptr) { + LOG_WARN("failed to cast subquery stmt to SelectStmt. "); + return RC::INVALID_ARGUMENT; + } + + // 子查询不能有超过一个列 + if (select_stmt->query_expressions_size() > 1) { + LOG_WARN("too many columns in subquery expression."); + return RC::TO_LONG_SUBQUERY_EXPR; + } + + // 将 select_stmt_ 指针设置为 select_stmt，使用 std::unique_ptr 来管理 + select_stmt_ = std::unique_ptr(select_stmt); + return RC::SUCCESS; +} + +RC SubQueryExpr::generate_logical_oper() +{ + RC rc = LogicalPlanGenerator::create(select_stmt_.get(), logical_oper_); + if (OB_FAIL(rc)) { + LOG_WARN("failed to generate logical operator for subquery. return %s", strrc(rc)); + return rc; + } + return RC::SUCCESS; +} + +RC SubQueryExpr::generate_physical_oper() +{ + RC rc = PhysicalPlanGenerator::create(*logical_oper_, physical_oper_); + if (OB_FAIL(rc)) { + LOG_WARN("failed to generate physical operator for subquery. return %s", strrc(rc)); + return rc; } return rc; -} \ No newline at end of file +} + +bool SubQueryExpr::one_row_ret() const { return res_query.size() <= 1; } + +// 子算子树的 open 和 close 逻辑由外部控制 +RC SubQueryExpr::open(Trx *trx, const Tuple &tuple) +{ + RC rc = RC::SUCCESS; + physical_oper_->set_parent_tuple(&tuple); + rc = physical_oper_->open(trx); + + return rc; +} + +RC SubQueryExpr::reset() +{ + visited_index = 0; + return RC::SUCCESS; +} + +bool SubQueryExpr::has_more_row(const Tuple &tuple) const +{ + physical_oper_->set_parent_tuple(&tuple); + return physical_oper_->next() != RC::RECORD_EOF; +} + +RC SubQueryExpr::get_value(const Tuple &tuple, Value &value) +{ + physical_oper_->set_parent_tuple(&tuple); + RC rc = physical_oper_->next(); + if (rc == RC::RECORD_EOF) { + // 先返回 null 类型的值，之后再完善具体选择的列类型 + // 如果已经成功执行过一次，结果不为空，那么这里的 value 不会被用到 + value.set_type(AttrType::NULLS); + value.set_null(true); + // 给调用者判断结果是否为空，而不是直接返回 RC::SUCCESS + return rc; + } else if (OB_FAIL(rc)) { + // 其他错误 + return rc; + } + + // 到这里确保有一条记录 + rc = physical_oper_->current_tuple()->cell_at(0, value); + if (OB_FAIL(rc)) { + return rc; + } + return RC::SUCCESS; +} + +RC SubQueryExpr::close() { return physical_oper_->close(); } + +RC SubQueryExpr::try_get_value(Value &value) const { return RC::UNIMPLEMENTED; } + +ExprType SubQueryExpr::type() const { return ExprType::SUBQUERY; } + +AttrType SubQueryExpr::value_type() const { return AttrType::UNDEFINED; } + +ListExpr::ListExpr(std::vector &&exprs) +{ + for (auto expr : exprs) { + exprs_.emplace_back(std::unique_ptr(expr)); + } + exprs.clear(); +} + +RC NormalFunctionExpr::type_from_string(const char *type_str, NormalFunctionType &type) +{ + check_type("typeof", NormalFunctionType::TYPEOF); + check_type("month", NormalFunctionType::MONTH); + check_type("year", NormalFunctionType::YEAR); + check_type("date_format", NormalFunctionType::DATE_FORMAT); + check_type("length", NormalFunctionType::LENGTH); + check_type("round", NormalFunctionType::ROUND); + check_type("l2_distance", NormalFunctionType::L2_DISTANCE); + check_type("cosine_distance", NormalFunctionType::COSINE_DISTANCE); + check_type("inner_product", NormalFunctionType::INNER_PRODUCT); + check_type("string_to_vector", NormalFunctionType::STRING_TO_VECTOR); + check_type("vector_to_string", NormalFunctionType::VECTOR_TO_STRING); + check_type("vector_dim", NormalFunctionType::VECTOR_DIM); + return RC::INVALID_ARGUMENT; +} + +RC NormalFunctionExpr::get_value(const Tuple &tuple, Value &result) +{ + vector args_values_; + for (auto &expr : args()) { + Value value; + RC rc = expr->get_value(tuple, value); + if (OB_FAIL(rc)) { + return rc; + } + args_values_.push_back(value); + } + switch (type_) { + case NormalFunctionType::LENGTH: return builtin::length(args_values_, result); + case NormalFunctionType::ROUND: return builtin::round(args_values_, result); + case NormalFunctionType::DATE_FORMAT: return builtin::date_format(args_values_, result); + case NormalFunctionType::STRING_TO_VECTOR: return builtin::string_to_vector(args_values_, result); + case NormalFunctionType::VECTOR_TO_STRING: return builtin::vector_to_string(args_values_, result); + case NormalFunctionType::VECTOR_DIM: return builtin::vector_dim(args_values_, result); + case NormalFunctionType::YEAR: return builtin::year(args_values_, result); + case NormalFunctionType::MONTH: return builtin::month(args_values_, result); + case NormalFunctionType::DAY: return builtin::day(args_values_, result); + case NormalFunctionType::L2_DISTANCE: return builtin::l2_distance(args_values_, result); + case NormalFunctionType::COSINE_DISTANCE: return builtin::cosine_distance(args_values_, result); + case NormalFunctionType::INNER_PRODUCT: return builtin::inner_product(args_values_, result); + case NormalFunctionType::TYPEOF: return builtin::_typeof(args_values_, result); + } + return RC::INTERNAL; +} + +RC NormalFunctionExpr::try_get_value(Value &result) const +{ + vector args_values_; + for (auto &expr : args()) { + Value value; + RC rc = expr->try_get_value(value); + if (OB_FAIL(rc)) { + return rc; + } + args_values_.push_back(value); + } + switch (type_) { + case NormalFunctionType::LENGTH: return builtin::length(args_values_, result); + case NormalFunctionType::ROUND: return builtin::round(args_values_, result); + case NormalFunctionType::DATE_FORMAT: return builtin::date_format(args_values_, result); + case NormalFunctionType::STRING_TO_VECTOR: return builtin::string_to_vector(args_values_, result); + case NormalFunctionType::VECTOR_TO_STRING: return builtin::vector_to_string(args_values_, result); + case NormalFunctionType::VECTOR_DIM: return builtin::vector_dim(args_values_, result); + case NormalFunctionType::YEAR: return builtin::year(args_values_, result); + case NormalFunctionType::MONTH: return builtin::month(args_values_, result); + case NormalFunctionType::DAY: return builtin::day(args_values_, result); + case NormalFunctionType::L2_DISTANCE: return builtin::l2_distance(args_values_, result); + case NormalFunctionType::COSINE_DISTANCE: return builtin::cosine_distance(args_values_, result); + case NormalFunctionType::INNER_PRODUCT: return builtin::inner_product(args_values_, result); + case NormalFunctionType::TYPEOF: return builtin::_typeof(args_values_, result); + } + return RC::INTERNAL; +} + +AttrType NormalFunctionExpr::value_type() const +{ + switch (type_) { + case NormalFunctionType::LENGTH: return AttrType::INTS; + case NormalFunctionType::ROUND: return AttrType::FLOATS; + case NormalFunctionType::DATE_FORMAT: return AttrType::CHARS; + case NormalFunctionType::STRING_TO_VECTOR: return AttrType::VECTORS; + case NormalFunctionType::VECTOR_TO_STRING: return AttrType::CHARS; + case NormalFunctionType::VECTOR_DIM: return AttrType::INTS; + case NormalFunctionType::YEAR: return AttrType::INTS; + case NormalFunctionType::MONTH: return AttrType::INTS; + case NormalFunctionType::DAY: return AttrType::INTS; + case NormalFunctionType::L2_DISTANCE: return AttrType::FLOATS; + case NormalFunctionType::COSINE_DISTANCE: return AttrType::FLOATS; + case NormalFunctionType::INNER_PRODUCT: return AttrType::FLOATS; + case NormalFunctionType::TYPEOF: return AttrType::CHARS; + } + return AttrType::UNDEFINED; +} diff --git a/src/observer/sql/expr/expression.h b/src/observer/sql/expr/expression.h index 7058fda4..bb689904 100644 --- a/src/observer/sql/expr/expression.h +++ b/src/observer/sql/expr/expression.h @@ -16,13 +16,18 @@ See the Mulan PSL v2 for more details. */ #include #include +#include #include "common/value.h" #include "storage/field/field.h" #include "sql/expr/aggregator.h" #include "storage/common/chunk.h" +#include "sql/builtin/builtin.h" class Tuple; +class SelectStmt; +class LogicalOperator; +class PhysicalOperator; /** * @defgroup Expression @@ -36,17 +41,19 @@ class Tuple; enum class ExprType { NONE, - STAR, ///< 星号，表示所有字段 - UNBOUND_FIELD, ///< 未绑定的字段，需要在resolver阶段解析为FieldExpr - UNBOUND_AGGREGATION, ///< 未绑定的聚合函数，需要在resolver阶段解析为AggregateExpr - - FIELD, ///< 字段。在实际执行时，根据行数据内容提取对应字段的值 - VALUE, ///< 常量值 - CAST, ///< 需要做类型转换的表达式 - COMPARISON, ///< 需要做比较的表达式 - CONJUNCTION, ///< 多个表达式使用同一种关系(AND或OR)来联结 - ARITHMETIC, ///< 算术运算 - AGGREGATION, ///< 聚合运算 + STAR, ///< 星号，表示所有字段 + UNBOUND_FIELD, ///< 未绑定的字段，需要在resolver阶段解析为FieldExpr + UNBOUND_FUNCTION, ///< 未绑定的聚合函数，需要在resolver阶段解析为AggregateFunctionExpr + FIELD, ///< 字段。在实际执行时，根据行数据内容提取对应字段的值 + VALUE, ///< 常量值 + CAST, ///< 需要做类型转换的表达式 + COMPARISON, ///< 需要做比较的表达式 + CONJUNCTION, ///< 多个表达式使用同一种关系(AND或OR)来联结 + ARITHMETIC, ///< 算术运算 + AGGREGATION, ///< 聚合运算 + NORMAL_FUNCTION, ///< 普通函数 + SUBQUERY, ///< 子查询 + EXPRLIST ///< 列表 }; /** @@ -75,7 +82,7 @@ class Expression /** * @brief 根据具体的tuple，来计算当前表达式的值。tuple有可能是一个具体某个表的行数据 */ - virtual RC get_value(const Tuple &tuple, Value &value) const = 0; + virtual RC get_value(const Tuple &tuple, Value &value) = 0; /** * @brief 在没有实际运行的情况下，也就是无法获取tuple的情况下，尝试获取表达式的值 @@ -109,7 +116,11 @@ class Expression * @brief 表达式的名字，比如是字段名称，或者用户在执行SQL语句时输入的内容 */ virtual const char *name() const { return name_.c_str(); } - virtual void set_name(std::string name) { name_ = name; } + virtual const char *alias() const { return alias_.c_str(); } + bool has_alias() const { return !alias_.empty(); } + virtual void set_name(std::string name) { name_ = std::move(name); } + virtual void set_alias(std::string alias) { alias_ = std::move(alias); } + virtual bool name_empty() { return name_.empty(); } /** * @brief 表达式在下层算子返回的 chunk 中的位置 @@ -122,6 +133,8 @@ class Expression */ virtual RC eval(Chunk &chunk, std::vector &select) { return RC::UNIMPLEMENTED; } + virtual RC reset() { return RC::SUCCESS; } + protected: /** * @brief 表达式在下层算子返回的 chunk 中的位置 @@ -133,19 +146,20 @@ class Expression private: std::string name_; + std::string alias_; }; class StarExpr : public Expression { public: StarExpr() : table_name_() {} - StarExpr(const char *table_name) : table_name_(table_name) {} - virtual ~StarExpr() = default; + explicit StarExpr(const char *table_name) : table_name_(table_name) {} + ~StarExpr() override = default; ExprType type() const override { return ExprType::STAR; } AttrType value_type() const override { return AttrType::UNDEFINED; } - RC get_value(const Tuple &tuple, Value &value) const override { return RC::UNIMPLEMENTED; } // 不需要实现 + RC get_value(const Tuple &tuple, Value &value) override { return RC::UNIMPLEMENTED; } // 不需要实现 const char *table_name() const { return table_name_.c_str(); } @@ -160,12 +174,12 @@ class UnboundFieldExpr : public Expression : table_name_(table_name), field_name_(field_name) {} - virtual ~UnboundFieldExpr() = default; + ~UnboundFieldExpr() override = default; ExprType type() const override { return ExprType::UNBOUND_FIELD; } AttrType value_type() const override { return AttrType::UNDEFINED; } - RC get_value(const Tuple &tuple, Value &value) const override { return RC::INTERNAL; } + RC get_value(const Tuple &tuple, Value &value) override { return RC::INTERNAL; } const char *table_name() const { return table_name_.c_str(); } const char *field_name() const { return field_name_.c_str(); } @@ -183,8 +197,8 @@ class FieldExpr : public Expression { public: FieldExpr() = default; - FieldExpr(const Table *table, const FieldMeta *field) : field_(table, field) {} - FieldExpr(const Field &field) : field_(field) {} + FieldExpr(const BaseTable *table, const FieldMeta *field) : field_(table, field) {} + explicit FieldExpr(const Field &field) : field_(field) {} virtual ~FieldExpr() = default; @@ -203,10 +217,13 @@ class FieldExpr : public Expression RC get_column(Chunk &chunk, Column &column) override; - RC get_value(const Tuple &tuple, Value &value) const override; + RC get_value(const Tuple &tuple, Value &value) override; + + void set_table_alias(std::string table_alias) { table_alias_ = std::move(table_alias); } private: - Field field_; + Field field_; + std::string table_alias_; }; /** @@ -219,11 +236,11 @@ class ValueExpr : public Expression ValueExpr() = default; explicit ValueExpr(const Value &value) : value_(value) {} - virtual ~ValueExpr() = default; + ~ValueExpr() override = default; bool equal(const Expression &other) const override; - RC get_value(const Tuple &tuple, Value &value) const override; + RC get_value(const Tuple &tuple, Value &value) override; RC get_column(Chunk &chunk, Column &column) override; RC try_get_value(Value &value) const override { @@ -254,7 +271,7 @@ class CastExpr : public Expression ExprType type() const override { return ExprType::CAST; } - RC get_value(const Tuple &tuple, Value &value) const override; + RC get_value(const Tuple &tuple, Value &value) override; RC try_get_value(Value &value) const override; @@ -277,11 +294,13 @@ class CastExpr : public Expression class ComparisonExpr : public Expression { public: + ComparisonExpr(CompOp comp, Expression *left, Expression *right); ComparisonExpr(CompOp comp, std::unique_ptr left, std::unique_ptr right); + virtual ~ComparisonExpr(); ExprType type() const override { return ExprType::COMPARISON; } - RC get_value(const Tuple &tuple, Value &value) const override; + RC get_value(const Tuple &tuple, Value &value) override; AttrType value_type() const override { return AttrType::BOOLEANS; } CompOp comp() const { return comp_; } @@ -332,11 +351,13 @@ class ConjunctionExpr : public Expression public: ConjunctionExpr(Type type, std::vector> &children); - virtual ~ConjunctionExpr() = default; + ConjunctionExpr(Type type, std::unique_ptr children); + ConjunctionExpr(Type type, Expression *left, Expression *right); + ~ConjunctionExpr() override = default; ExprType type() const override { return ExprType::CONJUNCTION; } AttrType value_type() const override { return AttrType::BOOLEANS; } - RC get_value(const Tuple &tuple, Value &value) const override; + RC get_value(const Tuple &tuple, Value &value) override; Type conjunction_type() const { return conjunction_type_; } @@ -380,7 +401,7 @@ class ArithmeticExpr : public Expression return 4; // sizeof(float) or sizeof(int) }; - RC get_value(const Tuple &tuple, Value &value) const override; + RC get_value(const Tuple &tuple, Value &value) override; RC get_column(Chunk &chunk, Column &column) override; @@ -405,42 +426,76 @@ class ArithmeticExpr : public Expression std::unique_ptr right_; }; -class UnboundAggregateExpr : public Expression +class UnboundFunctionExpr : public Expression { public: - UnboundAggregateExpr(const char *aggregate_name, Expression *child); - virtual ~UnboundAggregateExpr() = default; + UnboundFunctionExpr(const char *aggregate_name, std::vector> child); + virtual ~UnboundFunctionExpr() = default; - ExprType type() const override { return ExprType::UNBOUND_AGGREGATION; } + ExprType type() const override { return ExprType::UNBOUND_FUNCTION; } - const char *aggregate_name() const { return aggregate_name_.c_str(); } + const char *function_name() const { return function_name_.c_str(); } - std::unique_ptr &child() { return child_; } + string to_string() const + { + string str = function_name_ + "("; + for (size_t i = 0; i < args_.size(); i++) { + str += args_[i]->name(); + if (i < args_.size() - 1) { + str += ", "; + } + } + str += ")"; + return str; + } - RC get_value(const Tuple &tuple, Value &value) const override { return RC::INTERNAL; } - AttrType value_type() const override { return child_->value_type(); } + std::vector> &args() { return args_; } + const std::vector> &args() const { return args_; } + void set_args(std::vector> args) { args_ = std::move(args); } + + RC get_value(const Tuple &, Value &) override { return RC::INTERNAL; } + AttrType value_type() const override { return {}; } private: - std::string aggregate_name_; - std::unique_ptr child_; + std::string function_name_; + std::vector> args_; }; -class AggregateExpr : public Expression +class NormalFunctionExpr : public UnboundFunctionExpr { public: - enum class Type + NormalFunctionExpr( + NormalFunctionType type, const char *aggregate_name, std::vector> child) + : ::UnboundFunctionExpr(aggregate_name, std::move(child)), type_(type) + {} + + static RC type_from_string(const char *type_str, NormalFunctionType &type); + + ExprType type() const override { return ExprType::NORMAL_FUNCTION; } + + bool is_vector_distance_func() { - COUNT, - SUM, - AVG, - MAX, - MIN, - }; + return type_ == NormalFunctionType::L2_DISTANCE || type_ == NormalFunctionType::COSINE_DISTANCE || + type_ == NormalFunctionType::INNER_PRODUCT; + } + NormalFunctionType function_type() const { return type_; } + + AttrType value_type() const override; + + RC get_value(const Tuple &tuple, Value &value) override; + RC try_get_value(Value &value) const override; + +private: + NormalFunctionType type_; +}; + +class AggregateFunctionExpr : public Expression +{ public: - AggregateExpr(Type type, Expression *child); - AggregateExpr(Type type, std::unique_ptr child); - virtual ~AggregateExpr() = default; + AggregateFunctionExpr(AggregateFunctionType type, Expression *child); + AggregateFunctionExpr(AggregateFunctionType type, std::unique_ptr child); + virtual ~AggregateFunctionExpr() = default; bool equal(const Expression &other) const override; @@ -449,11 +504,11 @@ class AggregateExpr : public Expression AttrType value_type() const override { return child_->value_type(); } int value_length() const override { return child_->value_length(); } - RC get_value(const Tuple &tuple, Value &value) const override; + RC get_value(const Tuple &tuple, Value &value) override; RC get_column(Chunk &chunk, Column &column) override; - Type aggregate_type() const { return aggregate_type_; } + AggregateFunctionType aggregate_type() const { return aggregate_type_; } std::unique_ptr &child() { return child_; } @@ -461,10 +516,83 @@ class AggregateExpr : public Expression std::unique_ptr create_aggregator() const; -public: - static RC type_from_string(const char *type_str, Type &type); + static RC type_from_string(const char *type_str, AggregateFunctionType &type); private: - Type aggregate_type_; + AggregateFunctionType aggregate_type_; std::unique_ptr child_; -}; \ No newline at end of file +}; + +class SubQueryExpr : public Expression +{ +public: + explicit SubQueryExpr(SelectSqlNode &select_node); + virtual ~SubQueryExpr(); + + RC open(Trx *trx, const Tuple &tuple); + RC reset() override; + RC close(); + bool has_more_row(const Tuple &tuple) const; + + RC get_value(const Tuple &tuple, Value &value) override; + + RC try_get_value(Value &value) const override; + + ExprType type() const override; + + AttrType value_type() const override; + + std::unique_ptr deep_copy() const; + + RC generate_select_stmt(Db *db, const std::unordered_map &tables); + + RC generate_logical_oper(); + RC generate_physical_oper(); + + size_t res_nums() const { return res_query.size(); } + + bool one_row_ret() const; + +private: + SelectSqlNode &sql_node_; + std::unique_ptr select_stmt_; + std::unique_ptr logical_oper_; + std::unique_ptr physical_oper_; + + std::vector res_query; + size_t visited_index = 0; +}; + +class ListExpr : public Expression +{ +public: + explicit ListExpr(std::vector &&exprs); + explicit ListExpr(std::vector> &&exprs) : exprs_(std::move(exprs)) {} + virtual ~ListExpr() = default; + + RC reset() override + { + cur_idx_ = 0; + return RC::SUCCESS; + } + + RC get_value(const Tuple &tuple, Value &value) override + { + if (cur_idx_ >= exprs_.size()) { + return RC::RECORD_EOF; + } + return exprs_[cur_idx_++]->get_value(tuple, value); // 移除const_cast + } + + RC try_get_value(Value &value) const override { return RC::UNIMPLEMENTED; } + + ExprType type() const override { return ExprType::EXPRLIST; } + + AttrType value_type() const override { return AttrType::UNDEFINED; } + + std::vector> &get_list() { return exprs_; } + +private: + size_t cur_idx_ = 0; + std::vector> exprs_; +}; diff --git a/src/observer/sql/expr/expression_iterator.cpp b/src/observer/sql/expr/expression_iterator.cpp index ca8cd6d5..cdbdd1f7 100644 --- a/src/observer/sql/expr/expression_iterator.cpp +++ b/src/observer/sql/expr/expression_iterator.cpp @@ -13,34 +13,35 @@ See the Mulan PSL v2 for more details. */ // #include "sql/expr/expression_iterator.h" + +#include #include "sql/expr/expression.h" #include "common/log/log.h" using namespace std; -RC ExpressionIterator::iterate_child_expr(Expression &expr, function &)> callback) +RC ExpressionIterator::iterate_child_expr(Expression &expr, const function &)> &callback) { RC rc = RC::SUCCESS; switch (expr.type()) { case ExprType::CAST: { - auto &cast_expr = static_cast(expr); - rc = callback(cast_expr.child()); + auto &cast_expr = dynamic_cast(expr); + rc = callback(cast_expr.child()); } break; case ExprType::COMPARISON: { - - auto &comparison_expr = static_cast(expr); - rc = callback(comparison_expr.left()); - + auto &comparison_expr = dynamic_cast(expr); + rc = callback(comparison_expr.left()); if (OB_SUCC(rc)) { rc = callback(comparison_expr.right()); + } else { + return rc; } - } break; case ExprType::CONJUNCTION: { - auto &conjunction_expr = static_cast(expr); + auto &conjunction_expr = dynamic_cast(expr); for (auto &child : conjunction_expr.children()) { rc = callback(child); if (OB_FAIL(rc)) { @@ -50,19 +51,177 @@ RC ExpressionIterator::iterate_child_expr(Expression &expr, function(expr); - rc = callback(arithmetic_expr.left()); - if (OB_SUCC(rc)) { + auto &arithmetic_expr = dynamic_cast(expr); + rc = callback(arithmetic_expr.left()); + if (OB_SUCC(rc) && arithmetic_expr.right()) { rc = callback(arithmetic_expr.right()); } } break; case ExprType::AGGREGATION: { - auto &aggregate_expr = static_cast(expr); - rc = callback(aggregate_expr.child()); + auto &aggregate_expr = dynamic_cast(expr); + rc = callback(aggregate_expr.child()); + } break; + + case ExprType::NORMAL_FUNCTION: { + auto &normal_function_expr = dynamic_cast(expr); + for (auto &child_expr : normal_function_expr.args()) { + rc = callback(child_expr); + } + } break; + + case ExprType::NONE: + case ExprType::STAR: + case ExprType::UNBOUND_FIELD: + case ExprType::FIELD: + case ExprType::VALUE: { + // Do nothing + } break; + + default: { + ASSERT(false, "Unknown expression type"); + } + } + + return rc; +} + +RC ExpressionIterator::condition_iterate_expr(std::unique_ptr &expr) +{ + RC rc = RC::SUCCESS; + + switch (expr->type()) { + case ExprType::CAST: { + + } break; + + case ExprType::COMPARISON: { + auto *comp_expr = dynamic_cast(expr.get()); + auto &left = comp_expr->left(); + auto &right = comp_expr->right(); + + if (right->type() == ExprType::EXPRLIST) { + if (comp_expr->comp() != IN_OP && comp_expr->comp() != NOT_IN_OP) { + return RC::UNSUPPORTED; + } + } + + if (left->value_type() != AttrType::NULLS && right->value_type() != AttrType::NULLS) { + if (left->value_type() != right->value_type()) { + auto left_to_right_cost = Value::implicit_cast_cost(left->value_type(), right->value_type()); + auto right_to_left_cost = Value::implicit_cast_cost(right->value_type(), left->value_type()); + + if (right->type() == ExprType::SUBQUERY || right->type() == ExprType::EXPRLIST || + left->type() == ExprType::SUBQUERY || left->type() == ExprType::EXPRLIST) { + // 暂时在这里不做处理 + return RC::SUCCESS; + } else if (left_to_right_cost <= right_to_left_cost && left_to_right_cost != INT32_MAX) { + ExprType left_type = left->type(); + auto cast_expr = make_unique(std::move(left), right->value_type()); + + if (left_type == ExprType::VALUE) { + Value left_val; + rc = cast_expr->try_get_value(left_val); + if (OB_FAIL(rc)) { + LOG_WARN("failed to get value from left child", strrc(rc)); + return rc; + } + left = make_unique(left_val); + } else { + left = std::move(cast_expr); + } + } else if (right_to_left_cost < left_to_right_cost && right_to_left_cost != INT32_MAX) { + ExprType right_type = right->type(); + auto cast_expr = make_unique(std::move(right), left->value_type()); + + if (right_type == ExprType::VALUE) { + Value right_val; + rc = cast_expr->try_get_value(right_val); + if (OB_FAIL(rc)) { + LOG_WARN("failed to get value from right child", strrc(rc)); + return rc; + } + right = make_unique(right_val); + } else { + right = std::move(cast_expr); + } + } else { + rc = RC::UNSUPPORTED; + LOG_WARN("unsupported cast from %s to %s", + attr_type_to_string(left->value_type()), + attr_type_to_string(right->value_type())); + return rc; + } + } + } + } break; + + case ExprType::CONJUNCTION: { + auto conjunction_expr = dynamic_cast(expr.get()); + for (auto &child : conjunction_expr->children()) { + rc = condition_iterate_expr(child); + if (OB_FAIL(rc)) { + break; + } + } + } break; + + case ExprType::ARITHMETIC: + case ExprType::AGGREGATION: + case ExprType::NORMAL_FUNCTION: + case ExprType::NONE: + case ExprType::STAR: + case ExprType::UNBOUND_FIELD: + case ExprType::FIELD: + case ExprType::VALUE: { + // Do nothing + } break; + + default: { + ASSERT(false, "Unknown expression type"); + } + } + + return rc; +} + +RC ExpressionIterator::having_condition_iterate_expr( + std::unique_ptr &expr, vector &bound_expressions) +{ + RC rc = RC::SUCCESS; + + switch (expr->type()) { + case ExprType::COMPARISON: { + auto *comp_expr = dynamic_cast(expr.get()); + auto &left = comp_expr->left(); + auto &right = comp_expr->right(); + rc = having_condition_iterate_expr(left, bound_expressions); + if (OB_FAIL(rc)) { + break; + } + rc = having_condition_iterate_expr(right, bound_expressions); + if (OB_FAIL(rc)) { + break; + } + + } break; + + case ExprType::CONJUNCTION: { + auto conjunction_expr = dynamic_cast(expr.get()); + for (auto &child : conjunction_expr->children()) { + rc = having_condition_iterate_expr(child, bound_expressions); + if (OB_FAIL(rc)) { + break; + } + } + } break; + case ExprType::AGGREGATION: { + bound_expressions.push_back(expr.get()); } break; + case ExprType::CAST: + case ExprType::ARITHMETIC: + case ExprType::NORMAL_FUNCTION: case ExprType::NONE: case ExprType::STAR: case ExprType::UNBOUND_FIELD: @@ -77,4 +236,4 @@ RC ExpressionIterator::iterate_child_expr(Expression &expr, function &)> callback); -}; \ No newline at end of file + static RC iterate_child_expr(Expression &expr, const std::function &)> &callback); + static RC condition_iterate_expr(std::unique_ptr &expr); + static RC having_condition_iterate_expr( + std::unique_ptr &expr, std::vector &bound_expressions); +}; diff --git a/src/observer/sql/expr/expression_tuple.h b/src/observer/sql/expr/expression_tuple.h index 21f34fdb..cbd67303 100644 --- a/src/observer/sql/expr/expression_tuple.h +++ b/src/observer/sql/expr/expression_tuple.h @@ -29,6 +29,8 @@ class ExpressionTuple : public Tuple void set_tuple(const Tuple *tuple) { child_tuple_ = tuple; } + const Tuple *child_tuple() { return child_tuple_; } + int cell_num() const override { return static_cast(expressions_.size()); } RC cell_at(int index, Value &cell) const override @@ -85,7 +87,16 @@ class ExpressionTuple : public Tuple return rc; } + Tuple *copy() const override + { + auto copy = new ExpressionTuple(expressions_); + if (child_tuple_) { + copy->child_tuple_ = child_tuple_->copy(); + } + return copy; + } + private: const std::vector &expressions_; const Tuple *child_tuple_ = nullptr; -}; \ No newline at end of file +}; diff --git a/src/observer/sql/expr/tuple.h b/src/observer/sql/expr/tuple.h index 0215602a..219fe0bc 100644 --- a/src/observer/sql/expr/tuple.h +++ b/src/observer/sql/expr/tuple.h @@ -88,6 +88,8 @@ class Tuple virtual RC spec_at(int index, TupleCellSpec &spec) const = 0; + virtual Tuple *copy() const = 0; + /** * @brief 根据cell的描述，获取cell的值 * @@ -152,6 +154,17 @@ class Tuple result = 0; return rc; } + + void reset() { base_rids_.clear(); } + + void set_base_rids(std::vector> &base_rids) { base_rids_ = std::move(base_rids); } + + std::vector> &base_rids() { return base_rids_; } + + void append_base_rids(BaseTable *base_table, RID rid) { base_rids_.emplace_back(base_table, rid); } + +private: + std::vector> base_rids_; }; /** @@ -163,7 +176,7 @@ class RowTuple : public Tuple { public: RowTuple() = default; - virtual ~RowTuple() + ~RowTuple() override { for (FieldExpr *spec : speces_) { delete spec; @@ -173,7 +186,7 @@ class RowTuple : public Tuple void set_record(Record *record) { this->record_ = record; } - void set_schema(const Table *table, const std::vector *fields) + void set_schema(const BaseTable *table, const std::vector *fields) { table_ = table; // fix:join当中会多次调用右表的open,open当中会调用set_scheme，从而导致tuple当中会存储 @@ -188,6 +201,8 @@ class RowTuple : public Tuple } } + void set_table_alias(std::string &alias) { table_alias_ = std::move(alias); } + int cell_num() const override { return speces_.size(); } RC cell_at(int index, Value &cell) const override @@ -200,7 +215,19 @@ class RowTuple : public Tuple FieldExpr *field_expr = speces_[index]; const FieldMeta *field_meta = field_expr->field().meta(); cell.set_type(field_meta->type()); - cell.set_data(this->record_->data() + field_meta->offset(), field_meta->len()); + if (field_meta->nullable()) { + bool is_null = this->record_->data()[field_meta->offset() + field_meta->len() - 1] == '1'; + if (is_null) { + cell.set_null(is_null); + } else { + // 如果是字符型 null 值，这里虽然安全拷贝了 0 个数据，但因为 own_data，在发生拷贝构造时又由 length 0 + // 而没有初始化指针，导致内存越界 + cell.set_data(this->record_->data() + field_meta->offset(), field_meta->len() - 1); + } + } else { + cell.set_data(this->record_->data() + field_meta->offset(), field_meta->len()); + cell.set_null(false); + } return RC::SUCCESS; } @@ -215,6 +242,14 @@ class RowTuple : public Tuple { const char *table_name = spec.table_name(); const char *field_name = spec.field_name(); + const char *alias = spec.alias(); + + if (!common::is_blank(alias)) { + if (0 != strcmp(alias, table_alias_.c_str())) { + return RC::NOTFOUND; + } + } + if (0 != strcmp(table_name, table_->name())) { return RC::NOTFOUND; } @@ -245,10 +280,23 @@ class RowTuple : public Tuple const Record &record() const { return *record_; } + Tuple *copy() const override + { + RowTuple *copy = new RowTuple(); + for (auto &spec_ : speces_) { + copy->speces_.push_back(new FieldExpr(*spec_)); + } + copy->record_ = new Record(record_->clone()); + copy->table_ = table_; + copy->table_alias_ = table_alias_; + return copy; + } + private: Record *record_ = nullptr; - const Table *table_ = nullptr; + const BaseTable *table_ = nullptr; std::vector speces_; + std::string table_alias_; }; /** @@ -352,7 +400,7 @@ class ValueListTuple : public Tuple const int size = static_cast(specs_.size()); for (int i = 0; i < size; i++) { - if (specs_[i].equals(spec)) { + if (specs_[i] == spec) { cell = cells_[i]; return RC::SUCCESS; } @@ -382,6 +430,14 @@ class ValueListTuple : public Tuple return RC::SUCCESS; } + Tuple *copy() const override + { + auto copy = new ValueListTuple; + copy->cells_ = cells_; + copy->specs_ = specs_; + return copy; + } + private: std::vector cells_; std::vector specs_; @@ -402,6 +458,10 @@ class JoinedTuple : public Tuple void set_left(Tuple *left) { left_ = left; } void set_right(Tuple *right) { right_ = right; } + std::vector> &left_base_rids() { return left_->base_rids(); } + + std::vector> &right_base_rids() { return right_->base_rids(); } + int cell_num() const override { return left_->cell_num() + right_->cell_num(); } RC cell_at(int index, Value &value) const override @@ -442,7 +502,68 @@ class JoinedTuple : public Tuple return right_->find_cell(spec, value); } + Tuple *copy() const override + { + auto copy = new JoinedTuple; + if (left_) { + copy->left_ = left_->copy(); + } + if (right_) { + copy->right_ = right_->copy(); + } + return copy; + } + private: Tuple *left_ = nullptr; Tuple *right_ = nullptr; }; + +/** + * @brief 一些常量值组成的Tuple,用于 orderby 算子中 + * @ingroup Tuple + */ +class SplicedTuple : public Tuple +{ +public: + SplicedTuple() = default; + virtual ~SplicedTuple() = default; + + void set_cells(const std::vector &cells) { cells_ = cells; } + + int cell_num() const override { return cells_.size(); } + + RC cell_at(int index, Value &cell) const override + { + if (index < 0 || index >= cell_num()) { + return RC::NOTFOUND; + } + + cell = cells_[index]; + return RC::SUCCESS; + } + + RC find_cell(const TupleCellSpec &spec, Value &cell) const override + { + assert(false); + return RC::INTERNAL; + } + + RC init(const std::vector &exprs) + { + exprs_ = exprs; + return RC::SUCCESS; + } + + RC spec_at(int index, TupleCellSpec &spec) const override + { + assert(false); + return RC::INTERNAL; + } + + std::vector &exprs() { return exprs_; } + +private: + std::vector cells_; + std::vector exprs_; +}; diff --git a/src/observer/sql/expr/tuple_cell.cpp b/src/observer/sql/expr/tuple_cell.cpp index 4646643d..65a8c136 100644 --- a/src/observer/sql/expr/tuple_cell.cpp +++ b/src/observer/sql/expr/tuple_cell.cpp @@ -27,13 +27,8 @@ TupleCellSpec::TupleCellSpec(const char *table_name, const char *field_name, con } if (alias) { alias_ = alias; - } else { - if (table_name_.empty()) { - alias_ = field_name_; - } else { - alias_ = table_name_ + "." + field_name_; - } } + init_hash(); } TupleCellSpec::TupleCellSpec(const char *alias) @@ -41,7 +36,7 @@ TupleCellSpec::TupleCellSpec(const char *alias) if (alias) { alias_ = alias; } + init_hash(); } -TupleCellSpec::TupleCellSpec(const string &alias) : alias_(alias) -{} +TupleCellSpec::TupleCellSpec(const string &alias) : alias_(alias) { init_hash(); } diff --git a/src/observer/sql/expr/tuple_cell.h b/src/observer/sql/expr/tuple_cell.h index b9ba210e..78632aaf 100644 --- a/src/observer/sql/expr/tuple_cell.h +++ b/src/observer/sql/expr/tuple_cell.h @@ -29,13 +29,30 @@ class TupleCellSpec final const char *field_name() const { return field_name_.c_str(); } const char *alias() const { return alias_.c_str(); } - bool equals(const TupleCellSpec &other) const + bool alias_empty() const { return alias_.empty(); } + bool table_field_empty() const { return table_name_.empty() && field_name_.empty(); } + + bool operator==(const TupleCellSpec &other) const { - return table_name_ == other.table_name_ && field_name_ == other.field_name_ && alias_ == other.alias_; + return (!alias_empty() && !other.alias_empty() && alias_hash_ == other.alias_hash_ && alias_ == other.alias_) || + (!table_field_empty() && !other.table_field_empty() && table_name_hash_ == other.table_name_hash_ && + field_name_hash_ == other.field_name_hash_ && table_name_ == other.table_name_ && + field_name_ == other.field_name_); } private: std::string table_name_; std::string field_name_; std::string alias_; + + std::size_t table_name_hash_; + std::size_t field_name_hash_; + std::size_t alias_hash_; + + void init_hash() + { + table_name_hash_ = std::hash()(table_name_); + field_name_hash_ = std::hash()(field_name_); + alias_hash_ = std::hash()(alias_); + } }; diff --git a/src/observer/sql/operator/aggregate_vec_physical_operator.cpp b/src/observer/sql/operator/aggregate_vec_physical_operator.cpp index 710c6472..2419fb72 100644 --- a/src/observer/sql/operator/aggregate_vec_physical_operator.cpp +++ b/src/observer/sql/operator/aggregate_vec_physical_operator.cpp @@ -24,7 +24,7 @@ AggregateVecPhysicalOperator::AggregateVecPhysicalOperator(vector value_expressions_.reserve(aggregate_expressions_.size()); ranges::for_each(aggregate_expressions_, [this](Expression *expr) { - auto * aggregate_expr = static_cast(expr); + auto *aggregate_expr = static_cast(expr); Expression *child_expr = aggregate_expr->child().get(); ASSERT(child_expr != nullptr, "aggregation expression must have a child expression"); value_expressions_.emplace_back(child_expr); @@ -33,9 +33,9 @@ AggregateVecPhysicalOperator::AggregateVecPhysicalOperator(vector for (size_t i = 0; i < aggregate_expressions_.size(); i++) { auto &expr = aggregate_expressions_[i]; ASSERT(expr->type() == ExprType::AGGREGATION, "expected an aggregation expression"); - auto *aggregate_expr = static_cast(expr); + auto *aggregate_expr = static_cast(expr); - if (aggregate_expr->aggregate_type() == AggregateExpr::Type::SUM) { + if (aggregate_expr->aggregate_type() == AggregateFunctionType::SUM) { if (aggregate_expr->value_type() == AttrType::INTS) { void *aggr_value = malloc(sizeof(SumState)); ((SumState *)aggr_value)->value = 0; @@ -69,8 +69,8 @@ RC AggregateVecPhysicalOperator::open(Trx *trx) Column column; value_expressions_[aggr_idx]->get_column(chunk_, column); ASSERT(aggregate_expressions_[aggr_idx]->type() == ExprType::AGGREGATION, "expect aggregate expression"); - auto *aggregate_expr = static_cast(aggregate_expressions_[aggr_idx]); - if (aggregate_expr->aggregate_type() == AggregateExpr::Type::SUM) { + auto *aggregate_expr = static_cast(aggregate_expressions_[aggr_idx]); + if (aggregate_expr->aggregate_type() == AggregateFunctionType::SUM) { if (aggregate_expr->value_type() == AttrType::INTS) { update_aggregate_state, int>(aggr_values_.at(aggr_idx), column); } else if (aggregate_expr->value_type() == AttrType::FLOATS) { @@ -94,7 +94,7 @@ template void AggregateVecPhysicalOperator::update_aggregate_state(void *state, const Column &column) { STATE *state_ptr = reinterpret_cast(state); - T * data = (T *)column.data(); + T *data = (T *)column.data(); state_ptr->update(data, column.count()); } diff --git a/src/observer/sql/operator/calc_physical_operator.h b/src/observer/sql/operator/calc_physical_operator.h index d7002c82..272e693c 100644 --- a/src/observer/sql/operator/calc_physical_operator.h +++ b/src/observer/sql/operator/calc_physical_operator.h @@ -62,7 +62,11 @@ class CalcPhysicalOperator : public PhysicalOperator RC tuple_schema(TupleSchema &schema) const override { for (const std::unique_ptr &expression : expressions_) { - schema.append_cell(expression->name()); + if (expression->has_alias()) { + schema.append_cell(expression->alias()); + } else { + schema.append_cell(expression->name()); + } } return RC::SUCCESS; } diff --git a/src/observer/sql/operator/delete_logical_operator.cpp b/src/observer/sql/operator/delete_logical_operator.cpp index 5fc03f34..ed1c803c 100644 --- a/src/observer/sql/operator/delete_logical_operator.cpp +++ b/src/observer/sql/operator/delete_logical_operator.cpp @@ -14,4 +14,4 @@ See the Mulan PSL v2 for more details. */ #include "sql/operator/delete_logical_operator.h" -DeleteLogicalOperator::DeleteLogicalOperator(Table *table) : table_(table) {} +DeleteLogicalOperator::DeleteLogicalOperator(BaseTable *table) : table_(table) {} diff --git a/src/observer/sql/operator/delete_logical_operator.h b/src/observer/sql/operator/delete_logical_operator.h index c32ce662..73ddd521 100644 --- a/src/observer/sql/operator/delete_logical_operator.h +++ b/src/observer/sql/operator/delete_logical_operator.h @@ -23,12 +23,12 @@ See the Mulan PSL v2 for more details. */ class DeleteLogicalOperator : public LogicalOperator { public: - DeleteLogicalOperator(Table *table); + DeleteLogicalOperator(BaseTable *table); virtual ~DeleteLogicalOperator() = default; LogicalOperatorType type() const override { return LogicalOperatorType::DELETE; } - Table *table() const { return table_; } + BaseTable *table() const { return table_; } private: - Table *table_ = nullptr; + BaseTable *table_ = nullptr; }; diff --git a/src/observer/sql/operator/delete_physical_operator.cpp b/src/observer/sql/operator/delete_physical_operator.cpp index aaf34dfe..016d08b4 100644 --- a/src/observer/sql/operator/delete_physical_operator.cpp +++ b/src/observer/sql/operator/delete_physical_operator.cpp @@ -42,6 +42,7 @@ RC DeletePhysicalOperator::open(Trx *trx) RowTuple *row_tuple = static_cast(tuple); Record &record = row_tuple->record(); + record.set_base_rids(tuple->base_rids()); records_.emplace_back(std::move(record)); } @@ -60,12 +61,6 @@ RC DeletePhysicalOperator::open(Trx *trx) return RC::SUCCESS; } -RC DeletePhysicalOperator::next() -{ - return RC::RECORD_EOF; -} +RC DeletePhysicalOperator::next() { return RC::RECORD_EOF; } -RC DeletePhysicalOperator::close() -{ - return RC::SUCCESS; -} +RC DeletePhysicalOperator::close() { return RC::SUCCESS; } diff --git a/src/observer/sql/operator/delete_physical_operator.h b/src/observer/sql/operator/delete_physical_operator.h index d3c00fbb..12e8899f 100644 --- a/src/observer/sql/operator/delete_physical_operator.h +++ b/src/observer/sql/operator/delete_physical_operator.h @@ -26,7 +26,7 @@ class DeleteStmt; class DeletePhysicalOperator : public PhysicalOperator { public: - DeletePhysicalOperator(Table *table) : table_(table) {} + DeletePhysicalOperator(BaseTable *table) : table_(table) {} virtual ~DeletePhysicalOperator() = default; @@ -39,7 +39,7 @@ class DeletePhysicalOperator : public PhysicalOperator Tuple *current_tuple() override { return nullptr; } private: - Table *table_ = nullptr; + BaseTable *table_ = nullptr; Trx *trx_ = nullptr; std::vector records_; }; diff --git a/src/observer/sql/operator/explain_physical_operator.cpp b/src/observer/sql/operator/explain_physical_operator.cpp index da557370..0508e271 100644 --- a/src/observer/sql/operator/explain_physical_operator.cpp +++ b/src/observer/sql/operator/explain_physical_operator.cpp @@ -66,8 +66,8 @@ RC ExplainPhysicalOperator::next(Chunk &chunk) } generate_physical_plan(); - Value cell(physical_plan_.c_str()); - auto column = make_unique(); + Value cell(physical_plan_.c_str()); + auto column = make_unique(); column->init(cell); chunk.add_column(std::move(column), 0); return RC::SUCCESS; diff --git a/src/observer/sql/operator/group_by_logical_operator.cpp b/src/observer/sql/operator/group_by_logical_operator.cpp index c35bbf91..29e88fc7 100644 --- a/src/observer/sql/operator/group_by_logical_operator.cpp +++ b/src/observer/sql/operator/group_by_logical_operator.cpp @@ -20,9 +20,9 @@ See the Mulan PSL v2 for more details. */ using namespace std; -GroupByLogicalOperator::GroupByLogicalOperator(vector> &&group_by_exprs, - vector &&expressions) +GroupByLogicalOperator::GroupByLogicalOperator( + vector> &&group_by_exprs, vector &&expressions) { - group_by_expressions_ = std::move(group_by_exprs); + group_by_expressions_ = std::move(group_by_exprs); aggregate_expressions_ = std::move(expressions); } \ No newline at end of file diff --git a/src/observer/sql/operator/group_by_physical_operator.cpp b/src/observer/sql/operator/group_by_physical_operator.cpp index e28aea55..8828d86a 100644 --- a/src/observer/sql/operator/group_by_physical_operator.cpp +++ b/src/observer/sql/operator/group_by_physical_operator.cpp @@ -26,7 +26,7 @@ GroupByPhysicalOperator::GroupByPhysicalOperator(vector &&expressi aggregate_expressions_ = std::move(expressions); value_expressions_.reserve(aggregate_expressions_.size()); ranges::for_each(aggregate_expressions_, [this](Expression *expr) { - auto *aggregate_expr = static_cast(expr); + auto *aggregate_expr = static_cast(expr); Expression *child_expr = aggregate_expr->child().get(); ASSERT(child_expr != nullptr, "aggregate expression must have a child expression"); value_expressions_.emplace_back(child_expr); @@ -38,7 +38,7 @@ void GroupByPhysicalOperator::create_aggregator_list(AggregatorList &aggregator_ aggregator_list.clear(); aggregator_list.reserve(aggregate_expressions_.size()); ranges::for_each(aggregate_expressions_, [&aggregator_list](Expression *expr) { - auto *aggregate_expr = static_cast(expr); + auto *aggregate_expr = static_cast(expr); aggregator_list.emplace_back(aggregate_expr->create_aggregator()); }); } diff --git a/src/observer/sql/operator/group_by_vec_physical_operator.h b/src/observer/sql/operator/group_by_vec_physical_operator.h index e78469b4..3b2ee206 100644 --- a/src/observer/sql/operator/group_by_vec_physical_operator.h +++ b/src/observer/sql/operator/group_by_vec_physical_operator.h @@ -21,7 +21,7 @@ class GroupByVecPhysicalOperator : public PhysicalOperator { public: GroupByVecPhysicalOperator( - std::vector> &&group_by_exprs, std::vector &&expressions){}; + std::vector> &&group_by_exprs, std::vector &&expressions) {}; virtual ~GroupByVecPhysicalOperator() = default; diff --git a/src/observer/sql/operator/hash_group_by_physical_operator.cpp b/src/observer/sql/operator/hash_group_by_physical_operator.cpp index f408c0ec..dbda6008 100644 --- a/src/observer/sql/operator/hash_group_by_physical_operator.cpp +++ b/src/observer/sql/operator/hash_group_by_physical_operator.cpp @@ -23,8 +23,7 @@ using namespace common; HashGroupByPhysicalOperator::HashGroupByPhysicalOperator( vector> &&group_by_exprs, vector &&expressions) : GroupByPhysicalOperator(std::move(expressions)), group_by_exprs_(std::move(group_by_exprs)) -{ -} +{} RC HashGroupByPhysicalOperator::open(Trx *trx) { @@ -61,7 +60,7 @@ RC HashGroupByPhysicalOperator::open(Trx *trx) // 计算聚合值 GroupValueType &group_value = get<1>(*found_group); - rc = aggregate(get<0>(group_value), group_value_expression_tuple); + rc = aggregate(get<0>(group_value), group_value_expression_tuple); if (OB_FAIL(rc)) { LOG_WARN("failed to aggregate values. rc=%s", strrc(rc)); return rc; @@ -80,7 +79,7 @@ RC HashGroupByPhysicalOperator::open(Trx *trx) // 得到最终聚合后的值 for (GroupType &group : groups_) { GroupValueType &group_value = get<1>(group); - rc = evaluate(group_value); + rc = evaluate(group_value); if (OB_FAIL(rc)) { LOG_WARN("failed to evaluate group value. rc=%s", strrc(rc)); return rc; @@ -170,8 +169,8 @@ RC HashGroupByPhysicalOperator::find_group(const Tuple &child_tuple, GroupType * CompositeTuple composite_tuple; composite_tuple.add_tuple(make_unique(std::move(child_tuple_to_value))); - groups_.emplace_back(std::move(group_by_evaluated_tuple), - GroupValueType(std::move(aggregator_list), std::move(composite_tuple))); + groups_.emplace_back( + std::move(group_by_evaluated_tuple), GroupValueType(std::move(aggregator_list), std::move(composite_tuple))); found_group = &groups_.back(); } diff --git a/src/observer/sql/operator/index_scan_physical_operator.cpp b/src/observer/sql/operator/index_scan_physical_operator.cpp index 51bed88f..fd4a8966 100644 --- a/src/observer/sql/operator/index_scan_physical_operator.cpp +++ b/src/observer/sql/operator/index_scan_physical_operator.cpp @@ -16,13 +16,9 @@ See the Mulan PSL v2 for more details. */ #include "storage/index/index.h" #include "storage/trx/trx.h" -IndexScanPhysicalOperator::IndexScanPhysicalOperator(Table *table, Index *index, ReadWriteMode mode, const Value *left_value, - bool left_inclusive, const Value *right_value, bool right_inclusive) - : table_(table), - index_(index), - mode_(mode), - left_inclusive_(left_inclusive), - right_inclusive_(right_inclusive) +IndexScanPhysicalOperator::IndexScanPhysicalOperator(Table *table, Index *index, ReadWriteMode mode, + const Value *left_value, bool left_inclusive, const Value *right_value, bool right_inclusive) + : table_(table), index_(index), mode_(mode), left_inclusive_(left_inclusive), right_inclusive_(right_inclusive) { if (left_value) { left_value_ = *left_value; @@ -32,6 +28,19 @@ IndexScanPhysicalOperator::IndexScanPhysicalOperator(Table *table, Index *index, } } +IndexScanPhysicalOperator::IndexScanPhysicalOperator(Table *table, std::string table_alias, Index *index, + ReadWriteMode mode, const Value *left_value, bool left_inclusive, const Value *right_value, bool right_inclusive) + : table_(table), index_(index), mode_(mode), left_inclusive_(left_inclusive), right_inclusive_(right_inclusive) +{ + tuple_.set_table_alias(table_alias); + if (left_value) { + left_value_ = *left_value; + } + if (right_value) { + right_value_ = *right_value; + } +} + RC IndexScanPhysicalOperator::open(Trx *trx) { if (nullptr == table_ || nullptr == index_) { diff --git a/src/observer/sql/operator/index_scan_physical_operator.h b/src/observer/sql/operator/index_scan_physical_operator.h index 902c49f6..04347d19 100644 --- a/src/observer/sql/operator/index_scan_physical_operator.h +++ b/src/observer/sql/operator/index_scan_physical_operator.h @@ -25,10 +25,14 @@ See the Mulan PSL v2 for more details. */ class IndexScanPhysicalOperator : public PhysicalOperator { public: + // 实际上这个旧接口用不到了 IndexScanPhysicalOperator(Table *table, Index *index, ReadWriteMode mode, const Value *left_value, bool left_inclusive, const Value *right_value, bool right_inclusive); - virtual ~IndexScanPhysicalOperator() = default; + IndexScanPhysicalOperator(Table *table, std::string table_alias, Index *index, ReadWriteMode mode, + const Value *left_value, bool left_inclusive, const Value *right_value, bool right_inclusive); + + ~IndexScanPhysicalOperator() override = default; PhysicalOperatorType type() const override { return PhysicalOperatorType::INDEX_SCAN; } diff --git a/src/observer/sql/operator/insert_logical_operator.cpp b/src/observer/sql/operator/insert_logical_operator.cpp index 09f30e44..43774c1d 100644 --- a/src/observer/sql/operator/insert_logical_operator.cpp +++ b/src/observer/sql/operator/insert_logical_operator.cpp @@ -14,5 +14,6 @@ See the Mulan PSL v2 for more details. */ #include "sql/operator/insert_logical_operator.h" -InsertLogicalOperator::InsertLogicalOperator(Table *table, std::vector values) : table_(table), values_(values) +InsertLogicalOperator::InsertLogicalOperator(BaseTable *table, const std::vector> &values_list) + : table_(table), values_list_(values_list) {} diff --git a/src/observer/sql/operator/insert_logical_operator.h b/src/observer/sql/operator/insert_logical_operator.h index 18d9178a..ae2314e1 100644 --- a/src/observer/sql/operator/insert_logical_operator.h +++ b/src/observer/sql/operator/insert_logical_operator.h @@ -26,16 +26,15 @@ See the Mulan PSL v2 for more details. */ class InsertLogicalOperator : public LogicalOperator { public: - InsertLogicalOperator(Table *table, std::vector values); - virtual ~InsertLogicalOperator() = default; + InsertLogicalOperator(BaseTable *table, const std::vector> &); + ~InsertLogicalOperator() override = default; LogicalOperatorType type() const override { return LogicalOperatorType::INSERT; } - Table *table() const { return table_; } - const std::vector &values() const { return values_; } - std::vector &values() { return values_; } + BaseTable *table() const { return table_; } + const std::vector> &values_list() const { return values_list_; }; private: - Table *table_ = nullptr; - std::vector values_; -}; \ No newline at end of file + BaseTable *table_ = nullptr; + const std::vector> &values_list_; +}; diff --git a/src/observer/sql/operator/insert_physical_operator.cpp b/src/observer/sql/operator/insert_physical_operator.cpp index fd1bc908..77326416 100644 --- a/src/observer/sql/operator/insert_physical_operator.cpp +++ b/src/observer/sql/operator/insert_physical_operator.cpp @@ -17,25 +17,42 @@ See the Mulan PSL v2 for more details. */ #include "storage/table/table.h" #include "storage/trx/trx.h" -using namespace std; - -InsertPhysicalOperator::InsertPhysicalOperator(Table *table, vector &&values) - : table_(table), values_(std::move(values)) +InsertPhysicalOperator::InsertPhysicalOperator(BaseTable *table, const std::vector> &values_list) + : table_(table), values_list_(values_list) {} RC InsertPhysicalOperator::open(Trx *trx) { - Record record; - RC rc = table_->make_record(static_cast(values_.size()), values_.data(), record); - if (rc != RC::SUCCESS) { - LOG_WARN("failed to make record. rc=%s", strrc(rc)); - return rc; + RC rc; + std::vector records(values_list_.size()); + for (size_t i = 0; i < values_list_.size(); ++i) { + rc = table_->make_record(static_cast(values_list_[i].size()), values_list_[i].data(), records[i]); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to make record. rc=%s", strrc(rc)); + return rc; + } } - rc = trx->insert_record(table_, record); - if (rc != RC::SUCCESS) { - LOG_WARN("failed to insert record by transaction. rc=%s", strrc(rc)); + int size = static_cast(records.size()); + for (int i = 0; i < size; ++i) { + rc = trx->insert_record(table_, records[i]); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to insert record by transaction. rc=%s", strrc(rc)); + // 要么都插入成功，要么都不插入 + LOG_INFO("Rolling back previously inserted records up to index %d", i - 1); + while (--i >= 0) { + // 删除之前插入成功的 + RC rc2 = trx->delete_record(table_, records[i]); + if (rc2 != RC::SUCCESS) { + LOG_WARN("failed to delete record by transaction. rc=%s", strrc(rc2)); + } else { + LOG_INFO("Successfully deleted record at index %d", i); + } + } + return rc; + } } + return rc; } diff --git a/src/observer/sql/operator/insert_physical_operator.h b/src/observer/sql/operator/insert_physical_operator.h index ac47a538..25f8bd9b 100644 --- a/src/observer/sql/operator/insert_physical_operator.h +++ b/src/observer/sql/operator/insert_physical_operator.h @@ -27,9 +27,9 @@ class InsertStmt; class InsertPhysicalOperator : public PhysicalOperator { public: - InsertPhysicalOperator(Table *table, std::vector &&values); + InsertPhysicalOperator(BaseTable *table, const std::vector> &values_list); - virtual ~InsertPhysicalOperator() = default; + ~InsertPhysicalOperator() override = default; PhysicalOperatorType type() const override { return PhysicalOperatorType::INSERT; } @@ -40,6 +40,6 @@ class InsertPhysicalOperator : public PhysicalOperator Tuple *current_tuple() override { return nullptr; } private: - Table *table_ = nullptr; - std::vector values_; + BaseTable *table_ = nullptr; + const std::vector> &values_list_; }; diff --git a/src/observer/sql/operator/join_physical_operator.cpp b/src/observer/sql/operator/join_physical_operator.cpp index 0de38730..bd5c90c8 100644 --- a/src/observer/sql/operator/join_physical_operator.cpp +++ b/src/observer/sql/operator/join_physical_operator.cpp @@ -82,7 +82,16 @@ RC NestedLoopJoinPhysicalOperator::close() return rc; } -Tuple *NestedLoopJoinPhysicalOperator::current_tuple() { return &joined_tuple_; } +Tuple *NestedLoopJoinPhysicalOperator::current_tuple() +{ + // 拷贝 + auto left_base_rids = joined_tuple_.left_base_rids(); + auto right_base_rids = joined_tuple_.right_base_rids(); + left_base_rids.insert(left_base_rids.end(), right_base_rids.begin(), right_base_rids.end()); + joined_tuple_.set_base_rids(left_base_rids); + + return &joined_tuple_; +} RC NestedLoopJoinPhysicalOperator::left_next() { diff --git a/src/observer/sql/operator/limit_logical_operator.cpp b/src/observer/sql/operator/limit_logical_operator.cpp new file mode 100644 index 00000000..0ec55cae --- /dev/null +++ b/src/observer/sql/operator/limit_logical_operator.cpp @@ -0,0 +1,9 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ \ No newline at end of file diff --git a/src/observer/sql/operator/limit_logical_operator.h b/src/observer/sql/operator/limit_logical_operator.h new file mode 100644 index 00000000..cfee85d6 --- /dev/null +++ b/src/observer/sql/operator/limit_logical_operator.h @@ -0,0 +1,30 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +// +// Created by HuXin on 24-10-9. +// + +#pragma once + +#include "sql/operator/logical_operator.h" + +class LimitLogicalOperator : public LogicalOperator +{ +public: + LimitLogicalOperator(int limit) : limit_(limit) {} + + LogicalOperatorType type() const override { return LogicalOperatorType::LIMIT; } + + int limit() const { return limit_; } + +private: + int limit_; +}; \ No newline at end of file diff --git a/src/observer/sql/operator/limit_physical_operator.cpp b/src/observer/sql/operator/limit_physical_operator.cpp new file mode 100644 index 00000000..98afc1a4 --- /dev/null +++ b/src/observer/sql/operator/limit_physical_operator.cpp @@ -0,0 +1,26 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +#include "limit_physical_operator.h" + +RC LimitPhysicalOperator::open(Trx *trx) { return children_[0]->open(trx); } + +RC LimitPhysicalOperator::next() +{ + if (pos_ == limit_) { + return RC::RECORD_EOF; + } + pos_++; + return children_[0]->next(); +} + +RC LimitPhysicalOperator::close() { return children_[0]->close(); } + +Tuple *LimitPhysicalOperator::current_tuple() { return children_[0]->current_tuple(); } diff --git a/src/observer/sql/operator/limit_physical_operator.h b/src/observer/sql/operator/limit_physical_operator.h new file mode 100644 index 00000000..f0bddf43 --- /dev/null +++ b/src/observer/sql/operator/limit_physical_operator.h @@ -0,0 +1,40 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +// +// Created by HuXin on 24-10-9. +// + +#include "sql/operator/physical_operator.h" +#include "sql/expr/tuple.h" +#include +#include + +class LimitPhysicalOperator : public PhysicalOperator +{ +public: + LimitPhysicalOperator(int limit) : limit_(limit) {} + + virtual ~LimitPhysicalOperator() = default; + + PhysicalOperatorType type() const override { return PhysicalOperatorType::LIMIT; } + + int limit() const { return limit_; } + + RC open(Trx *trx) override; + RC next() override; + RC close() override; + + Tuple *current_tuple() override; + +private: + int pos_ = 0; + int limit_; +}; diff --git a/src/observer/sql/operator/logical_operator.cpp b/src/observer/sql/operator/logical_operator.cpp index d11d6fd1..28b6dfec 100644 --- a/src/observer/sql/operator/logical_operator.cpp +++ b/src/observer/sql/operator/logical_operator.cpp @@ -19,19 +19,12 @@ LogicalOperator::~LogicalOperator() {} void LogicalOperator::add_child(std::unique_ptr oper) { children_.emplace_back(std::move(oper)); } bool LogicalOperator::can_generate_vectorized_operator(const LogicalOperatorType &type) { - bool bool_ret = false; - switch (type) - { - case LogicalOperatorType::CALC: + bool bool_ret = false; + switch (type) { case LogicalOperatorType::DELETE: - case LogicalOperatorType::INSERT: - bool_ret = false; - break; - - default: - bool_ret = true; - break; - } - return bool_ret; -} + case LogicalOperatorType::INSERT: bool_ret = false; break; + default: bool_ret = true; break; + } + return bool_ret; +} diff --git a/src/observer/sql/operator/logical_operator.h b/src/observer/sql/operator/logical_operator.h index 32fb13cd..bc6a5bb5 100644 --- a/src/observer/sql/operator/logical_operator.h +++ b/src/observer/sql/operator/logical_operator.h @@ -39,8 +39,11 @@ enum class LogicalOperatorType JOIN, ///< 连接 INSERT, ///< 插入 DELETE, ///< 删除，删除可能会有子查询 + UPDATE, ///< 更新，更新可能会有子查询 EXPLAIN, ///< 查看执行计划 GROUP_BY, ///< 分组 + ORDER_BY, ///< 排序 + LIMIT, ///< 限制输出 }; /** diff --git a/src/observer/sql/operator/order_by_logical_operator.cpp b/src/observer/sql/operator/order_by_logical_operator.cpp new file mode 100644 index 00000000..0a57089f --- /dev/null +++ b/src/observer/sql/operator/order_by_logical_operator.cpp @@ -0,0 +1,15 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +// +// Created by HuXin on 24-10-9. +// + +#include "order_by_logical_operator.h" diff --git a/src/observer/sql/operator/order_by_logical_operator.h b/src/observer/sql/operator/order_by_logical_operator.h new file mode 100644 index 00000000..648c16c3 --- /dev/null +++ b/src/observer/sql/operator/order_by_logical_operator.h @@ -0,0 +1,34 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +// +// Created by HuXin on 24-10-9. +// + +#pragma once + +#include "sql/operator/logical_operator.h" + +/** + * @brief 逻辑算子 + * @ingroup LogicalOperator + */ +class OrderByLogicalOperator : public LogicalOperator +{ +public: + OrderByLogicalOperator(std::vector order_by) : order_by_(std::move(order_by)) {} + + LogicalOperatorType type() const override { return LogicalOperatorType::ORDER_BY; } + + std::vector &order_by() { return order_by_; } + +private: + std::vector order_by_; +}; \ No newline at end of file diff --git a/src/observer/sql/operator/order_by_physical_operator.cpp b/src/observer/sql/operator/order_by_physical_operator.cpp new file mode 100644 index 00000000..2ada8994 --- /dev/null +++ b/src/observer/sql/operator/order_by_physical_operator.cpp @@ -0,0 +1,97 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +// +// Created by HuXin on 24-10-9. +// + +#include "order_by_physical_operator.h" + +#include + +OrderByPhysicalOperator::OrderByPhysicalOperator(vector order_by) : order_by_(std::move(order_by)) +{ + order_and_field_line = order_list([this](const order_line &cells_a, const order_line &cells_b) -> bool { + auto order_size = order_by_.size(); + auto &order_line_a = cells_a.first; + auto &order_line_b = cells_b.first; + assert(order_line_a.size() == order_size); + assert(order_line_b.size() == order_size); + assert(order_by_.size() == order_size); + + for (size_t i = 0; i < order_size; i++) { + auto &a = order_line_a[i]; + auto &b = order_line_b[i]; + auto result = a.compare(b); + auto is_asc = order_by_[i].is_asc; + if (result < 0) { + // a < b + return !is_asc; + } else if (result > 0) { + // a > 0 + return is_asc; + } + } + + // order_line_a == order_line_b + return true; + }); +} + +RC OrderByPhysicalOperator::fetch_and_sort_tables() +{ + RC rc = RC::SUCCESS; + + while (RC::SUCCESS == (rc = children_[0]->next())) { + // 获取 order by 字段的 values + vector order_by_line; + for (auto &[expr, asc] : order_by_) { + Value cell; + rc = expr->get_value(*children_[0]->current_tuple(), cell); + if (OB_FAIL(rc)) { + return rc; + } + order_by_line.emplace_back(cell); + } + + order_and_field_line.emplace(order_by_line, children_[0]->current_tuple()->copy()); + } + + return RC::SUCCESS; +} + +RC OrderByPhysicalOperator::open(Trx *trx) +{ + RC rc = RC::SUCCESS; + if (children_.size() != 1) { + return RC::INTERNAL; + } + rc = children_[0]->open(trx); + if (OB_FAIL(rc)) { + return rc; + } + rc = fetch_and_sort_tables(); + return rc; +} + +RC OrderByPhysicalOperator::next() +{ + if (order_and_field_line.empty()) { + return RC::RECORD_EOF; + } + + tuple_ = order_and_field_line.top().second; + order_and_field_line.pop(); + return RC::SUCCESS; +} + +RC OrderByPhysicalOperator::close() { return children_[0]->close(); } + +Tuple *OrderByPhysicalOperator::current_tuple() { return tuple_; } diff --git a/src/observer/sql/operator/order_by_physical_operator.h b/src/observer/sql/operator/order_by_physical_operator.h new file mode 100644 index 00000000..47d3d47b --- /dev/null +++ b/src/observer/sql/operator/order_by_physical_operator.h @@ -0,0 +1,48 @@ +/* Copyright (c) 2021 OceanBase and/or its affiliates. All rights reserved. +miniob is licensed under Mulan PSL v2. +You can use this software according to the terms and conditions of the Mulan PSL v2. +You may obtain a copy of Mulan PSL v2 at: + http://license.coscl.org.cn/MulanPSL2 +THIS SOFTWARE IS PROVIDED ON AN "AS IS" BASIS, WITHOUT WARRANTIES OF ANY KIND, +EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, +MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. +See the Mulan PSL v2 for more details. */ + +// +// Created by HuXin on 24-10-9. +// + +#include "sql/operator/physical_operator.h" +#include "sql/expr/tuple.h" +#include "sql/expr/expression_tuple.h" +#include +#include + +class OrderByPhysicalOperator : public PhysicalOperator +{ +public: + OrderByPhysicalOperator(std::vector order_by); + + virtual ~OrderByPhysicalOperator() = default; + + PhysicalOperatorType type() const override { return PhysicalOperatorType::ORDER_BY; } + + std::vector &order_by() { return order_by_; } + + RC fetch_and_sort_tables(); + RC open(Trx *trx) override; + RC next() override; + RC close() override; + + Tuple *current_tuple() override; + +private: + std::vector order_by_; + + using order_line = pair, Tuple *>; + using order_func = std::function; + using order_list = std::priority_queue, order_func>; + order_list order_and_field_line; + + Tuple *tuple_ = nullptr; +}; diff --git a/src/observer/sql/operator/physical_operator.cpp b/src/observer/sql/operator/physical_operator.cpp index fa81906d..f846778e 100644 --- a/src/observer/sql/operator/physical_operator.cpp +++ b/src/observer/sql/operator/physical_operator.cpp @@ -19,11 +19,16 @@ std::string physical_operator_type_name(PhysicalOperatorType type) switch (type) { case PhysicalOperatorType::TABLE_SCAN: return "TABLE_SCAN"; case PhysicalOperatorType::INDEX_SCAN: return "INDEX_SCAN"; + case PhysicalOperatorType::VIEW_SCAN: return "VIEW_SCAN"; + case PhysicalOperatorType::VECTOR_INDEX_SCAN: return "VECTOR_INDEX_SCAN"; + case PhysicalOperatorType::LIMIT: return "LIMIT"; + case PhysicalOperatorType::ORDER_BY: return "ORDER_BY"; case PhysicalOperatorType::NESTED_LOOP_JOIN: return "NESTED_LOOP_JOIN"; case PhysicalOperatorType::EXPLAIN: return "EXPLAIN"; case PhysicalOperatorType::PREDICATE: return "PREDICATE"; case PhysicalOperatorType::INSERT: return "INSERT"; case PhysicalOperatorType::DELETE: return "DELETE"; + case PhysicalOperatorType::UPDATE: return "UPDATE"; case PhysicalOperatorType::PROJECT: return "PROJECT"; case PhysicalOperatorType::STRING_LIST: return "STRING_LIST"; case PhysicalOperatorType::HASH_GROUP_BY: return "HASH_GROUP_BY"; @@ -40,3 +45,10 @@ std::string physical_operator_type_name(PhysicalOperatorType type) std::string PhysicalOperator::name() const { return physical_operator_type_name(type()); } std::string PhysicalOperator::param() const { return ""; } +void PhysicalOperator::set_parent_tuple(const Tuple *tuple) +{ + parent_tuple_ = tuple; + for (auto &child : children_) { + child->set_parent_tuple(tuple); + } +} diff --git a/src/observer/sql/operator/physical_operator.h b/src/observer/sql/operator/physical_operator.h index a3b20ed3..30352523 100644 --- a/src/observer/sql/operator/physical_operator.h +++ b/src/observer/sql/operator/physical_operator.h @@ -40,6 +40,8 @@ enum class PhysicalOperatorType TABLE_SCAN, TABLE_SCAN_VEC, INDEX_SCAN, + VIEW_SCAN, + VECTOR_INDEX_SCAN, NESTED_LOOP_JOIN, EXPLAIN, PREDICATE, @@ -50,10 +52,13 @@ enum class PhysicalOperatorType STRING_LIST, DELETE, INSERT, + UPDATE, SCALAR_GROUP_BY, HASH_GROUP_BY, GROUP_BY_VEC, AGGREGATE_VEC, + ORDER_BY, + LIMIT, EXPR_VEC, }; @@ -86,9 +91,11 @@ class PhysicalOperator virtual RC tuple_schema(TupleSchema &schema) const { return RC::UNIMPLEMENTED; } void add_child(std::unique_ptr oper) { children_.emplace_back(std::move(oper)); } + void set_parent_tuple(const Tuple *tuple); std::vector> &children() { return children_; } protected: std::vector> children_; + const Tuple *parent_tuple_ = nullptr; // 不相关子查询的时候使用 }; diff --git a/src/observer/sql/operator/predicate_physical_operator.cpp b/src/observer/sql/operator/predicate_physical_operator.cpp index 61b56445..bbea61c0 100644 --- a/src/observer/sql/operator/predicate_physical_operator.cpp +++ b/src/observer/sql/operator/predicate_physical_operator.cpp @@ -38,6 +38,8 @@ RC PredicatePhysicalOperator::next() RC rc = RC::SUCCESS; PhysicalOperator *oper = children_.front().get(); + Tuple *tp = nullptr; + JoinedTuple jt; while (RC::SUCCESS == (rc = oper->next())) { Tuple *tuple = oper->current_tuple(); if (nullptr == tuple) { @@ -45,9 +47,16 @@ RC PredicatePhysicalOperator::next() LOG_WARN("failed to get tuple from operator"); break; } + if (parent_tuple_) { + jt.set_left(tuple); + jt.set_right(const_cast(parent_tuple_)); + tp = &jt; + } else { + tp = tuple; + } Value value; - rc = expression_->get_value(*tuple, value); + rc = expression_->get_value(*tp, value); if (rc != RC::SUCCESS) { return rc; } @@ -67,7 +76,4 @@ RC PredicatePhysicalOperator::close() Tuple *PredicatePhysicalOperator::current_tuple() { return children_[0]->current_tuple(); } -RC PredicatePhysicalOperator::tuple_schema(TupleSchema &schema) const -{ - return children_[0]->tuple_schema(schema); -} +RC PredicatePhysicalOperator::tuple_schema(TupleSchema &schema) const { return children_[0]->tuple_schema(schema); } diff --git a/src/observer/sql/operator/project_physical_operator.cpp b/src/observer/sql/operator/project_physical_operator.cpp index 9c06125b..048cb3bb 100644 --- a/src/observer/sql/operator/project_physical_operator.cpp +++ b/src/observer/sql/operator/project_physical_operator.cpp @@ -20,9 +20,8 @@ See the Mulan PSL v2 for more details. */ using namespace std; ProjectPhysicalOperator::ProjectPhysicalOperator(vector> &&expressions) - : expressions_(std::move(expressions)), tuple_(expressions_) -{ -} + : expressions_(std::move(expressions)), tuple_(expressions_) +{} RC ProjectPhysicalOperator::open(Trx *trx) { @@ -55,9 +54,12 @@ RC ProjectPhysicalOperator::close() } return RC::SUCCESS; } + Tuple *ProjectPhysicalOperator::current_tuple() { - tuple_.set_tuple(children_[0]->current_tuple()); + auto tuple = children_[0]->current_tuple(); + tuple_.set_base_rids(tuple->base_rids()); + tuple_.set_tuple(tuple); return &tuple_; } @@ -67,4 +69,4 @@ RC ProjectPhysicalOperator::tuple_schema(TupleSchema &schema) const schema.append_cell(expression->name()); } return RC::SUCCESS; -} \ No newline at end of file +} diff --git a/src/observer/sql/operator/scalar_group_by_physical_operator.cpp b/src/observer/sql/operator/scalar_group_by_physical_operator.cpp index b67e9f16..eb092bb5 100644 --- a/src/observer/sql/operator/scalar_group_by_physical_operator.cpp +++ b/src/observer/sql/operator/scalar_group_by_physical_operator.cpp @@ -65,7 +65,7 @@ RC ScalarGroupByPhysicalOperator::open(Trx *trx) composite_tuple.add_tuple(make_unique(std::move(child_tuple_to_value))); group_value_ = make_unique(std::move(aggregator_list), std::move(composite_tuple)); } - + rc = aggregate(get<0>(*group_value_), group_value_expression_tuple); if (OB_FAIL(rc)) { LOG_WARN("failed to aggregate values. rc=%s", strrc(rc)); @@ -93,7 +93,26 @@ RC ScalarGroupByPhysicalOperator::open(Trx *trx) RC ScalarGroupByPhysicalOperator::next() { + if (emitted_) { + return RC::RECORD_EOF; + } if (group_value_ == nullptr || emitted_) { + auto *aggregate_expr = static_cast(aggregate_expressions_[0]); + if (aggregate_expr->aggregate_type() == AggregateFunctionType::COUNT) { + Value val(0); + auto Vlist = make_unique(); + Vlist->set_cells(std::vector{val}); + TupleCellSpec spec(aggregate_expr->name()); + Vlist->set_names(std::vector{spec}); + CompositeTuple composite_tuple; + composite_tuple.add_tuple(std::move(Vlist)); + + AggregatorList aggregator_list; + group_value_ = make_unique(std::move(aggregator_list), std::move(composite_tuple)); + emitted_ = true; + + return RC::SUCCESS; + } return RC::RECORD_EOF; } diff --git a/src/observer/sql/operator/string_list_physical_operator.h b/src/observer/sql/operator/string_list_physical_operator.h index 86ffd3fc..2ed7ddfd 100644 --- a/src/observer/sql/operator/string_list_physical_operator.h +++ b/src/observer/sql/operator/string_list_physical_operator.h @@ -35,7 +35,7 @@ class StringListPhysicalOperator : public PhysicalOperator strings_.emplace_back(begin, end); } - void append(std::initializer_list init) { strings_.emplace_back(init); } + void append(const std::vector &init) { strings_.emplace_back(init); } template void append(const T &v) diff --git a/src/observer/sql/operator/table_get_logical_operator.cpp b/src/observer/sql/operator/table_get_logical_operator.cpp index 6933c0cd..672af036 100644 --- a/src/observer/sql/operator/table_get_logical_operator.cpp +++ b/src/observer/sql/operator/table_get_logical_operator.cpp @@ -14,9 +14,15 @@ See the Mulan PSL v2 for more details. */ #include "sql/operator/table_get_logical_operator.h" -TableGetLogicalOperator::TableGetLogicalOperator(Table *table, ReadWriteMode mode) +#include + +TableGetLogicalOperator::TableGetLogicalOperator(BaseTable *table, ReadWriteMode mode) : table_(table), mode_(mode) {} + +TableGetLogicalOperator::TableGetLogicalOperator(BaseTable *table, std::string table_alias, ReadWriteMode mode) : table_(table), mode_(mode) -{} +{ + table_alias_ = std::move(table_alias); +} void TableGetLogicalOperator::set_predicates(std::vector> &&exprs) { diff --git a/src/observer/sql/operator/table_get_logical_operator.h b/src/observer/sql/operator/table_get_logical_operator.h index 4551a371..4f173246 100644 --- a/src/observer/sql/operator/table_get_logical_operator.h +++ b/src/observer/sql/operator/table_get_logical_operator.h @@ -25,24 +25,44 @@ See the Mulan PSL v2 for more details. */ class TableGetLogicalOperator : public LogicalOperator { public: - TableGetLogicalOperator(Table *table, ReadWriteMode mode); - virtual ~TableGetLogicalOperator() = default; + TableGetLogicalOperator(BaseTable *table, ReadWriteMode mode); + TableGetLogicalOperator(BaseTable *table, std::string table_alias, ReadWriteMode mode); + ~TableGetLogicalOperator() override = default; LogicalOperatorType type() const override { return LogicalOperatorType::TABLE_GET; } - Table *table() const { return table_; } + BaseTable *table() const { return table_; } ReadWriteMode read_write_mode() const { return mode_; } void set_predicates(std::vector> &&exprs); auto predicates() -> std::vector> & { return predicates_; } + std::string &table_alias() { return table_alias_; } + + std::vector &base_vector() { return base_vector_; } + void set_base_vector(std::vector base_vector) { base_vector_ = std::move(base_vector); } + + size_t limit() const { return limit_; } + void set_limit(size_t limit) { limit_ = limit; } + + Index *index() const { return index_; } + void set_index(Index *index) { index_ = index; } + + bool is_vector_scan() { return !base_vector_.empty(); } + private: - Table *table_ = nullptr; - ReadWriteMode mode_ = ReadWriteMode::READ_WRITE; + BaseTable *table_ = nullptr; + std::string table_alias_; + ReadWriteMode mode_ = ReadWriteMode::READ_WRITE; // 与当前表相关的过滤操作，可以尝试在遍历数据时执行 // 这里的表达式都是比较简单的比较运算，并且左右两边都是取字段表达式或值表达式 // 不包含复杂的表达式运算，比如加减乘除、或者conjunction expression // 如果有多个表达式，他们的关系都是 AND std::vector> predicates_; + + // 向量索引参数 + Index *index_ = nullptr; + std::vector base_vector_; // 要比较的向量 + std::size_t limit_; // 输出多少个 }; diff --git a/src/observer/sql/operator/table_scan_physical_operator.cpp b/src/observer/sql/operator/table_scan_physical_operator.cpp index c7eabaaa..1c4179a3 100644 --- a/src/observer/sql/operator/table_scan_physical_operator.cpp +++ b/src/observer/sql/operator/table_scan_physical_operator.cpp @@ -35,7 +35,10 @@ RC TableScanPhysicalOperator::next() bool filter_result = false; while (OB_SUCC(rc = record_scanner_.next(current_record_))) { LOG_TRACE("got a record. rid=%s", current_record_.rid().to_string().c_str()); - + + // 存储记录来自哪个表和 rid + tuple_.reset(); + tuple_.append_base_rids(table_, current_record_.rid()); tuple_.set_record(¤t_record_); rc = filter(tuple_, filter_result); if (rc != RC::SUCCESS) { @@ -70,11 +73,19 @@ void TableScanPhysicalOperator::set_predicates(vector> && RC TableScanPhysicalOperator::filter(RowTuple &tuple, bool &result) { - RC rc = RC::SUCCESS; - Value value; + RC rc = RC::SUCCESS; + Value value; + Tuple *tp = &tuple; + JoinedTuple jt; + jt.set_left(&tuple); + jt.set_right(const_cast(parent_tuple_)); + if (parent_tuple_) { + tp = &jt; + } for (unique_ptr &expr : predicates_) { - rc = expr->get_value(tuple, value); + rc = expr->get_value(*tp, value); if (rc != RC::SUCCESS) { + close(); return rc; } diff --git a/src/observer/sql/operator/table_scan_physical_operator.h b/src/observer/sql/operator/table_scan_physical_operator.h index 206ecfd8..28678031 100644 --- a/src/observer/sql/operator/table_scan_physical_operator.h +++ b/src/observer/sql/operator/table_scan_physical_operator.h @@ -30,7 +30,12 @@ class TableScanPhysicalOperator : public PhysicalOperator public: TableScanPhysicalOperator(Table *table, ReadWriteMode mode) : table_(table), mode_(mode) {} - virtual ~TableScanPhysicalOperator() = default; + TableScanPhysicalOperator(Table *table, std::string alias, ReadWriteMode mode) : table_(table), mode_(mode) + { + tuple_.set_table_alias(alias); + } + + ~TableScanPhysicalOperator() override = default; std::string param() const override; diff --git a/src/observer/sql/operator/update_logical_operator.cpp b/src/observer/sql/operator/update_logical_operator.cpp new file mode 100644 index 00000000..397b810a --- /dev/null +++ b/src/observer/sql/operator/update_logical_operator.cpp @@ -0,0 +1,18 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/20 * + * @Description : UpdateLogicalOperator source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "update_logical_operator.h" + +UpdateLogicalOperator::UpdateLogicalOperator( + BaseTable *table, std::vector field_metas, std::vector> values) + : table_(table), field_metas_(std::move(field_metas)), values_(std::move(values)) +{} diff --git a/src/observer/sql/operator/update_logical_operator.h b/src/observer/sql/operator/update_logical_operator.h new file mode 100644 index 00000000..f9e4a6dc --- /dev/null +++ b/src/observer/sql/operator/update_logical_operator.h @@ -0,0 +1,37 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/20 * + * @Description : UpdateLogicalOperator header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "sql/operator/logical_operator.h" + +/** + * @brief 逻辑算子，用于执行udpate语句 + * @ingroup LogicalOperator + */ +class UpdateLogicalOperator : public LogicalOperator +{ +public: + explicit UpdateLogicalOperator( + BaseTable *table, std::vector field_metas, std::vector> values); + ~UpdateLogicalOperator() override = default; + + LogicalOperatorType type() const override { return LogicalOperatorType::UPDATE; } + BaseTable *table() const { return table_; } + std::vector &field_metas() { return field_metas_; } + std::vector> &values() { return values_; } + +private: + BaseTable *table_ = nullptr; + std::vector field_metas_; + std::vector> values_; +}; diff --git a/src/observer/sql/operator/update_physical_operator.cpp b/src/observer/sql/operator/update_physical_operator.cpp new file mode 100644 index 00000000..da52e69a --- /dev/null +++ b/src/observer/sql/operator/update_physical_operator.cpp @@ -0,0 +1,179 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/17 * + * @Description : UpdatePhysicalOperator source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "update_physical_operator.h" +#include "storage/trx/trx.h" + +RC UpdatePhysicalOperator::open(Trx *trx) +{ + if (children_.empty()) { + return RC::SUCCESS; + } + + std::unique_ptr &child = children_[0]; + + RC rc = child->open(trx); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to open child operator: %s", strrc(rc)); + return rc; + } + + trx_ = trx; + + while (OB_SUCC(rc = child->next())) { + Tuple *tuple = child->current_tuple(); + if (nullptr == tuple) { + LOG_WARN("failed to get current record: %s", strrc(rc)); + return rc; + } + + RowTuple *row_tuple = static_cast(tuple); + Record &record = row_tuple->record(); + record.set_base_rids(tuple->base_rids()); + records_.emplace_back(std::move(record)); + } + + child->close(); + + // 如果需要更新的记录为空，直接返回成功，即使 value 校验异常也应该返回成功 + if (records_.empty()) { + return RC::SUCCESS; + } + + // 得到真正的 value，并做校验 + RowTuple tuple; + Value value; + std::vector real_values(values_.size()); + SubQueryExpr *sub_query_expr = nullptr; + int size = static_cast(values_.size()); + for (int i = 0; i < size; ++i) { + auto &value_expr = values_[i]; + auto &field_meta = field_metas_[i]; + + if (value_expr->type() == ExprType::SUBQUERY) { + sub_query_expr = dynamic_cast(value_expr.get()); + rc = sub_query_expr->open(trx_, tuple); + if (OB_FAIL(rc)) { + LOG_ERROR("Failed to open subquery for field: %s", + field_meta.name()); + return rc; + } + } + + // 得到表达式的值 + rc = value_expr->get_value(tuple, value); + if (OB_FAIL(rc) && rc != RC::RECORD_EOF) { + LOG_ERROR("Failed to get value for field: %s", + field_meta.name()); + return rc; + } + + // 如果是子查询只能有一行一列 + if (sub_query_expr && sub_query_expr->has_more_row(tuple)) { + LOG_ERROR("Subquery returned more than one row for field: %s", + field_meta.name()); + return RC::SUBQUERY_RETURNED_MULTIPLE_ROWS; + } + + // 进行类型校验 + if (value.attr_type() != field_meta.type()) { + // 尝试转换，发生转换时不考虑数值溢出 + Value to_value; + // 更新不允许非目标类型的类型提升 + rc = Value::cast_to(value, field_meta.type(), to_value, false); + if (rc != RC::SUCCESS) { + LOG_ERROR("Schema field type mismatch and cast to failed. Field: %s, Expected Type: %s, Provided Type: %s, Length: %d", + field_meta.name(), + attr_type_to_string(field_meta.type()), + attr_type_to_string(value.attr_type()), value.length()); + return RC::SCHEMA_FIELD_TYPE_MISMATCH; + } + // 转换成功 + value = std::move(to_value); + } + + // 进行长度校验 + if (value.length() > field_meta.len() - field_meta.nullable()) { + LOG_ERROR("Value length exceeds maximum allowed length for field. Field: %s, Type: %s, Offset: %d, Length: %d, Max Length: %d", + field_meta.name(), + attr_type_to_string(field_meta.type()), + field_meta.offset(), + value.length(), + field_meta.len()); + return RC::VALUE_TOO_LONG; + } + + if (sub_query_expr) { + sub_query_expr->close(); + sub_query_expr = nullptr; + } + + real_values[i] = std::move(value); + } + + // 先收集记录再更新 + // 记录的有效性由事务来保证，如果事务不保证删除的有效性，那说明此事务类型不支持并发控制，比如VacuousTrx + Record new_record; + for (Record &old_record : records_) { + // rid 得手动拷贝 + new_record.set_rid(old_record.rid()); + new_record.copy_data(old_record.data(), old_record.len()); + for (size_t i = 0; i < field_metas_.size(); ++i) { + if (field_metas_[i].nullable()) { + auto null_offset = field_metas_[i].offset() + field_metas_[i].len() - 1; + if (real_values[i].is_null()) { + new_record.data()[null_offset] = '1'; + } else { + new_record.data()[null_offset] = 0; + } + } else if (real_values[i].is_null()) { + rollback(); + return RC::NOT_NULLABLE_VALUE; + } + // 只有非 null 值才需要拷贝数据，防止读到垃圾数据 + if (!real_values[i].is_null()) { + rc = new_record.set_field(field_metas_[i].offset(), field_metas_[i].len(), real_values[i]); + if (OB_FAIL(rc)) { + LOG_ERROR("failed to set field: %s", strrc(rc)); + return rc; + } + } + } + auto rollback_old_record = old_record.clone(); + auto rollback_new_record = new_record.clone(); + rc = trx_->update_record(table_, old_record, new_record); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to update record: %s", strrc(rc)); + rollback(); + return rc; + } else { + log_records.emplace_back(rollback_old_record, rollback_new_record); + } + } + + return RC::SUCCESS; +} + +void UpdatePhysicalOperator::rollback() +{ + RC rc; + for (auto &[rollback_old_record, rollback_new_record] : log_records) { + rc = trx_->update_record(table_, rollback_new_record, rollback_old_record); + if (OB_FAIL(rc)) { + LOG_WARN("failed to update record: %s", strrc(rc)); + } + } +} + +RC UpdatePhysicalOperator::next() { return RC::RECORD_EOF; } + +RC UpdatePhysicalOperator::close() { return RC::SUCCESS; } diff --git a/src/observer/sql/operator/update_physical_operator.h b/src/observer/sql/operator/update_physical_operator.h new file mode 100644 index 00000000..865b240c --- /dev/null +++ b/src/observer/sql/operator/update_physical_operator.h @@ -0,0 +1,51 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/17 * + * @Description : UpdatePhysicalOperator header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "sql/operator/physical_operator.h" + +class Trx; +class UpdateStmt; + +/** + * @brief 物理算子，更新 + * @ingroup PhysicalOperator + */ +class UpdatePhysicalOperator : public PhysicalOperator +{ +public: + UpdatePhysicalOperator( + BaseTable *table, std::vector field_metas, std::vector> values) + : table_(table), field_metas_(std::move(field_metas)), values_(std::move(values)) + {} + + ~UpdatePhysicalOperator() override = default; + + PhysicalOperatorType type() const override { return PhysicalOperatorType::UPDATE; } + + RC open(Trx *trx) override; + RC next() override; + RC close() override; + + Tuple *current_tuple() override { return nullptr; } + +private: + void rollback(); + + Trx *trx_ = nullptr; + BaseTable *table_ = nullptr; + std::vector field_metas_; + std::vector> values_; + std::vector records_; + vector> log_records; +}; diff --git a/src/observer/sql/operator/vector_scan_physical_operator.cpp b/src/observer/sql/operator/vector_scan_physical_operator.cpp new file mode 100644 index 00000000..18d87a1f --- /dev/null +++ b/src/observer/sql/operator/vector_scan_physical_operator.cpp @@ -0,0 +1,128 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/18 * + * @Description : VectorScanPhysicalOperator source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "storage/trx/trx.h" +#include "vector_scan_physical_operator.h" + +VectorScanPhysicalOperator::VectorScanPhysicalOperator(Table *table, std::string table_alias, Index *index, + std::vector base_vector, size_t limit, ReadWriteMode mode) + : table_(table), + index_(dynamic_cast(index)), + base_vector_(std::move(base_vector)), + limit_(limit), + mode_(mode) +{ + tuple_.set_table_alias(table_alias); +} + +RC VectorScanPhysicalOperator::open(Trx *trx) +{ + if (nullptr == table_ || nullptr == index_) { + return RC::INTERNAL; + } + + rids_ = index_->ann_search(base_vector_, limit_); + + record_handler_ = table_->record_handler(); + if (nullptr == record_handler_) { + LOG_WARN("invalid record handler"); + return RC::INTERNAL; + } + + tuple_.set_schema(table_, table_->table_meta().field_metas()); + + trx_ = trx; + return RC::SUCCESS; +} + +RC VectorScanPhysicalOperator::next() +{ + RC rc = RC::SUCCESS; + + if (cnt_ >= rids_.size()) { + return RC::RECORD_EOF; + } + + bool filter_result = false; + while (cnt_ < rids_.size()) { + auto rid = rids_[cnt_++]; + + rc = record_handler_->get_record(rid, current_record_); + if (OB_FAIL(rc)) { + LOG_TRACE("failed to get record. rid=%s, rc=%s", rid.to_string().c_str(), strrc(rc)); + return rc; + } + + LOG_TRACE("got a record. rid=%s", rid.to_string().c_str()); + + tuple_.set_record(¤t_record_); + rc = filter(tuple_, filter_result); + if (OB_FAIL(rc)) { + LOG_TRACE("failed to filter record. rc=%s", strrc(rc)); + return rc; + } + + if (!filter_result) { + LOG_TRACE("record filtered"); + continue; + } + + rc = trx_->visit_record(table_, current_record_, mode_); + if (rc == RC::RECORD_INVISIBLE) { + LOG_TRACE("record invisible"); + continue; + } else { + return rc; + } + } + + return rc; +} + +RC VectorScanPhysicalOperator::close() { return RC::SUCCESS; } + +Tuple *VectorScanPhysicalOperator::current_tuple() +{ + tuple_.set_record(¤t_record_); + return &tuple_; +} + +void VectorScanPhysicalOperator::set_predicates(std::vector> &&exprs) +{ + predicates_ = std::move(exprs); +} + +RC VectorScanPhysicalOperator::filter(RowTuple &tuple, bool &result) +{ + RC rc = RC::SUCCESS; + Value value; + for (std::unique_ptr &expr : predicates_) { + rc = expr->get_value(tuple, value); + if (rc != RC::SUCCESS) { + return rc; + } + + bool tmp_result = value.get_boolean(); + if (!tmp_result) { + result = false; + return rc; + } + } + + result = true; + return rc; +} + +std::string VectorScanPhysicalOperator::param() const +{ + return std::string(index_->index_meta().name()) + " ON " + table_->name(); +} diff --git a/src/observer/sql/operator/vector_scan_physical_operator.h b/src/observer/sql/operator/vector_scan_physical_operator.h new file mode 100644 index 00000000..2b453ba0 --- /dev/null +++ b/src/observer/sql/operator/vector_scan_physical_operator.h @@ -0,0 +1,66 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/18 * + * @Description : VectorScanPhysicalOperator header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "sql/expr/tuple.h" +#include "sql/operator/physical_operator.h" +#include "storage/record/record_manager.h" +#include "storage/index/ivfflat_index.h" + +/** + * @brief 向量索引扫描物理算子 + * @ingroup PhysicalOperator + */ +class VectorScanPhysicalOperator : public PhysicalOperator +{ +public: + // 向量索引的构造方法 + VectorScanPhysicalOperator(Table *table, std::string table_alias, Index *index, std::vector base_vector, + size_t limit, ReadWriteMode mode); + + ~VectorScanPhysicalOperator() override = default; + + PhysicalOperatorType type() const override { return PhysicalOperatorType::VECTOR_INDEX_SCAN; } + + std::string param() const override; + + RC open(Trx *trx) override; + RC next() override; + RC close() override; + + Tuple *current_tuple() override; + + void set_predicates(std::vector> &&exprs); + +private: + // 与TableScanPhysicalOperator代码相同，可以优化 + RC filter(RowTuple &tuple, bool &result); + +private: + Trx *trx_ = nullptr; + Table *table_ = nullptr; + IvfflatIndex *index_ = nullptr; + std::vector rids_; // 向量近似扫描结果 + RecordFileHandler *record_handler_ = nullptr; + + const std::vector base_vector_; + size_t limit_; + size_t cnt_{0}; + + ReadWriteMode mode_ = ReadWriteMode::READ_WRITE; + + Record current_record_{}; + RowTuple tuple_; + + std::vector> predicates_; +}; diff --git a/src/observer/sql/operator/view_scan_physical_operator.cpp b/src/observer/sql/operator/view_scan_physical_operator.cpp new file mode 100644 index 00000000..04ffe336 --- /dev/null +++ b/src/observer/sql/operator/view_scan_physical_operator.cpp @@ -0,0 +1,186 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/13 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "sql/operator/view_scan_physical_operator.h" +#include "event/sql_debug.h" +#include "storage/table/table.h" +#include "sql/expr/expression_tuple.h" +#include "storage/table/view.h" +#include "sql/stmt/select_stmt.h" +#include "sql/optimizer/logical_plan_generator.h" +#include "sql/optimizer/physical_plan_generator.h" + +RC ViewScanPhysicalOperator::init() +{ + RC rc = RC::SUCCESS; + + auto select_sql = view_->select_sql(); + + ParsedSqlResult parsed_sql_result; + parse(select_sql.c_str(), &parsed_sql_result); + + if (parsed_sql_result.sql_nodes().empty()) { + LOG_ERROR("Parsing failed: No SQL nodes found"); + return RC::INTERNAL; + } + + if (parsed_sql_result.sql_nodes().size() > 1) { + LOG_WARN("got multi sql commands but only 1 will be handled"); + } + + std::unique_ptr sql_node = std::move(parsed_sql_result.sql_nodes().front()); + if (sql_node->flag == SCF_ERROR) { + LOG_ERROR("Syntax error in SQL"); + return RC::SQL_SYNTAX; + } + if (sql_node->flag != SCF_SELECT) { + LOG_ERROR("Unexpected SQL command type. Expected SELECT, got flag: %d", sql_node->flag); + return RC::INTERNAL; + } + + Stmt *stmt = nullptr; + rc = SelectStmt::create(view_->db(), sql_node->selection, stmt); + if (rc != RC::SUCCESS && rc != RC::UNIMPLEMENTED) { + LOG_WARN("Failed to create select statement. rc=%d:%s", rc, strrc(rc)); + return rc; + } + + auto select_stmt = dynamic_cast(stmt); + std::unique_ptr logical_oper = nullptr; + LogicalPlanGenerator::create(select_stmt, logical_oper); + + if (logical_oper == nullptr) { + LOG_ERROR("Failed to create logical plan"); + return RC::INTERNAL; + } + + std::unique_ptr physical_oper = nullptr; + PhysicalPlanGenerator::create(*logical_oper, physical_oper); + + if (physical_oper == nullptr) { + LOG_ERROR("Failed to create physical plan"); + return RC::INTERNAL; + } + + select_expr_ = std::move(physical_oper); + + return RC::SUCCESS; +} + +RC ViewScanPhysicalOperator::open(Trx *trx) +{ + RC rc = RC::SUCCESS; + + if (select_expr_ == nullptr) { + rc = init(); + if (OB_FAIL(rc)) { + return rc; + } + } + + rc = select_expr_->open(trx); + if (rc == RC::SUCCESS) { + tuple_.set_schema(view_, view_->table_meta().field_metas()); + } + trx_ = trx; + return rc; +} + +RC ViewScanPhysicalOperator::next() +{ + RC rc = RC::SUCCESS; + + bool filter_result = false; + while (OB_SUCC(rc = next_record())) { + LOG_TRACE("got a record. rid=%s", current_record_.rid().to_string().c_str()); + + tuple_.set_record(¤t_record_); + rc = filter(tuple_, filter_result); + if (rc != RC::SUCCESS) { + LOG_TRACE("record filtered failed=%s", strrc(rc)); + return rc; + } + + if (filter_result) { + sql_debug("get a tuple: %s", tuple_.to_string().c_str()); + break; + } else { + sql_debug("a tuple is filtered: %s", tuple_.to_string().c_str()); + } + } + return rc; +} + +RC ViewScanPhysicalOperator::close() { return select_expr_->close(); } + +Tuple *ViewScanPhysicalOperator::current_tuple() { return &tuple_; } + +string ViewScanPhysicalOperator::param() const { return view_->name(); } + +void ViewScanPhysicalOperator::set_predicates(vector> &&exprs) +{ + predicates_ = std::move(exprs); +} + +RC ViewScanPhysicalOperator::filter(RowTuple &tuple, bool &result) +{ + RC rc = RC::SUCCESS; + Value value; + Tuple *tp = &tuple; + JoinedTuple jt; + jt.set_left(&tuple); + jt.set_right(const_cast(parent_tuple_)); + if (parent_tuple_) { + tp = &jt; + } + for (unique_ptr &expr : predicates_) { + rc = expr->get_value(*tp, value); + if (rc != RC::SUCCESS) { + close(); + return rc; + } + + bool tmp_result = value.get_boolean(); + if (!tmp_result) { + result = false; + return rc; + } + } + + result = true; + return rc; +} + +RC ViewScanPhysicalOperator::next_record() +{ + RC rc = select_expr_->next(); + if (OB_FAIL(rc)) { + return rc; + } + + auto tuple = select_expr_->current_tuple(); + tuple_.set_base_rids(tuple->base_rids()); + // 构造一个 view 的 record + int cell_num = tuple->cell_num(); + std::vector values(cell_num); + for (int i = 0; i < cell_num; i++) { + Value cell; + rc = tuple->cell_at(i, cell); + if (OB_FAIL(rc)) { + return rc; + } + values[i] = std::move(cell); + } + + rc = view_->make_record(static_cast(values.size()), values.data(), current_record_); + return rc; +} diff --git a/src/observer/sql/operator/view_scan_physical_operator.h b/src/observer/sql/operator/view_scan_physical_operator.h new file mode 100644 index 00000000..80f1d226 --- /dev/null +++ b/src/observer/sql/operator/view_scan_physical_operator.h @@ -0,0 +1,59 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/13 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "common/rc.h" +#include "sql/operator/physical_operator.h" +#include "storage/record/record_manager.h" +#include "sql/expr/expression_tuple.h" +#include "storage/table/view.h" + +/** + * @brief 视图扫描物理算子 + * @ingroup PhysicalOperator + */ +class ViewScanPhysicalOperator : public PhysicalOperator +{ +public: + ViewScanPhysicalOperator(View *view, std::string alias) : view_(view) { tuple_.set_table_alias(alias); } + + ~ViewScanPhysicalOperator() override = default; + + std::string param() const override; + + PhysicalOperatorType type() const override { return PhysicalOperatorType::VIEW_SCAN; } + + RC open(Trx *trx) override; + RC next() override; + RC close() override; + + Tuple *current_tuple() override; + + void set_predicates(std::vector> &&exprs); + +private: + RC filter(RowTuple &tuple, bool &result); + + RC init(); + + // 把查询的结果转为当前视图的记录，就相当于从视图表中取出来一个记录 + RC next_record(); + +private: + View *view_ = nullptr; + Trx *trx_ = nullptr; + std::unique_ptr select_expr_ = nullptr; + Record current_record_; + RowTuple tuple_; + std::vector> predicates_; // TODO chang predicate to table tuple filter +}; diff --git a/src/observer/sql/optimizer/conjunction_simplification_rule.cpp b/src/observer/sql/optimizer/conjunction_simplification_rule.cpp index d60a1b55..f410d5e7 100644 --- a/src/observer/sql/optimizer/conjunction_simplification_rule.cpp +++ b/src/observer/sql/optimizer/conjunction_simplification_rule.cpp @@ -32,17 +32,17 @@ RC ConjunctionSimplificationRule::rewrite(std::unique_ptr &expr, boo return rc; } - change_made = false; - auto conjunction_expr = static_cast(expr.get()); + change_made = false; + auto conjunction_expr = static_cast(expr.get()); - std::vector> &child_exprs = conjunction_expr->children(); + std::vector> &child_exprs = conjunction_expr->children(); // 先看看有没有能够直接去掉的表达式。比如AND时恒为true的表达式可以删除 // 或者是否可以直接计算出当前表达式的值。比如AND时，如果有一个表达式为false，那么整个表达式就是false for (auto iter = child_exprs.begin(); iter != child_exprs.end();) { bool constant_value = false; - rc = try_to_get_bool_constant(*iter, constant_value); + rc = try_to_get_bool_constant(*iter, constant_value); if (rc != RC::SUCCESS) { rc = RC::SUCCESS; ++iter; diff --git a/src/observer/sql/optimizer/expression_rewriter.cpp b/src/observer/sql/optimizer/expression_rewriter.cpp index b0c6d245..7b9a0da0 100644 --- a/src/observer/sql/optimizer/expression_rewriter.cpp +++ b/src/observer/sql/optimizer/expression_rewriter.cpp @@ -69,7 +69,7 @@ RC ExpressionRewriter::rewrite_expression(unique_ptr &expr, bool &ch for (unique_ptr &rule : expr_rewrite_rules_) { bool sub_change_made = false; - rc = rule->rewrite(expr, sub_change_made); + rc = rule->rewrite(expr, sub_change_made); if (sub_change_made && !change_made) { change_made = true; } @@ -91,25 +91,25 @@ RC ExpressionRewriter::rewrite_expression(unique_ptr &expr, bool &ch case ExprType::CAST: { unique_ptr &child_expr = (static_cast(expr.get()))->child(); - rc = rewrite_expression(child_expr, change_made); + rc = rewrite_expression(child_expr, change_made); } break; case ExprType::COMPARISON: { - auto comparison_expr = static_cast(expr.get()); + auto comparison_expr = static_cast(expr.get()); - unique_ptr &left_expr = comparison_expr->left(); - unique_ptr &right_expr = comparison_expr->right(); + unique_ptr &left_expr = comparison_expr->left(); + unique_ptr &right_expr = comparison_expr->right(); bool left_change_made = false; - rc = rewrite_expression(left_expr, left_change_made); + rc = rewrite_expression(left_expr, left_change_made); if (rc != RC::SUCCESS) { return rc; } bool right_change_made = false; - rc = rewrite_expression(right_expr, right_change_made); + rc = rewrite_expression(right_expr, right_change_made); if (rc != RC::SUCCESS) { return rc; } @@ -120,13 +120,13 @@ RC ExpressionRewriter::rewrite_expression(unique_ptr &expr, bool &ch } break; case ExprType::CONJUNCTION: { - auto conjunction_expr = static_cast(expr.get()); + auto conjunction_expr = static_cast(expr.get()); - vector> &children = conjunction_expr->children(); + vector> &children = conjunction_expr->children(); for (unique_ptr &child_expr : children) { bool sub_change_made = false; - rc = rewrite_expression(child_expr, sub_change_made); + rc = rewrite_expression(child_expr, sub_change_made); if (rc != RC::SUCCESS) { LOG_WARN("failed to rewriter conjunction sub expression. rc=%s", strrc(rc)); diff --git a/src/observer/sql/optimizer/logical_plan_generator.cpp b/src/observer/sql/optimizer/logical_plan_generator.cpp index 8c418b32..516c8493 100644 --- a/src/observer/sql/optimizer/logical_plan_generator.cpp +++ b/src/observer/sql/optimizer/logical_plan_generator.cpp @@ -22,10 +22,12 @@ See the Mulan PSL v2 for more details. */ #include "sql/operator/insert_logical_operator.h" #include "sql/operator/join_logical_operator.h" #include "sql/operator/logical_operator.h" +#include "sql/operator/order_by_logical_operator.h" #include "sql/operator/predicate_logical_operator.h" #include "sql/operator/project_logical_operator.h" #include "sql/operator/table_get_logical_operator.h" #include "sql/operator/group_by_logical_operator.h" +#include "sql/operator/limit_logical_operator.h" #include "sql/stmt/calc_stmt.h" #include "sql/stmt/delete_stmt.h" @@ -37,6 +39,9 @@ See the Mulan PSL v2 for more details. */ #include "sql/expr/expression_iterator.h" +#include "sql/stmt/update_stmt.h" +#include "sql/operator/update_logical_operator.h" + using namespace std; using namespace common; @@ -68,6 +73,12 @@ RC LogicalPlanGenerator::create(Stmt *stmt, unique_ptr &logical rc = create_plan(delete_stmt, logical_operator); } break; + case StmtType::UPDATE: { + UpdateStmt *update_stmt = static_cast(stmt); + + rc = create_plan(update_stmt, logical_operator); + } break; + case StmtType::EXPLAIN: { ExplainStmt *explain_stmt = static_cast(stmt); @@ -93,10 +104,11 @@ RC LogicalPlanGenerator::create_plan(SelectStmt *select_stmt, unique_ptr table_oper(nullptr); last_oper = &table_oper; - const std::vector &tables = select_stmt->tables(); - for (Table *table : tables) { - - unique_ptr table_get_oper(new TableGetLogicalOperator(table, ReadWriteMode::READ_ONLY)); + const std::vector &tables = select_stmt->tables(); + const std::vector &alias = select_stmt->tables_alias(); + for (int i = 0; i < tables.size(); ++i) { + unique_ptr table_get_oper( + new TableGetLogicalOperator(tables[i], alias[i], ReadWriteMode::READ_ONLY)); if (table_oper == nullptr) { table_oper = std::move(table_get_oper); } else { @@ -138,6 +150,44 @@ RC LogicalPlanGenerator::create_plan(SelectStmt *select_stmt, unique_ptr having_predicate_oper; + + rc = create_plan(select_stmt->having_filter_stmt(), having_predicate_oper); + if (OB_FAIL(rc)) { + LOG_WARN("failed to create predicate logical plan. rc=%s", strrc(rc)); + return rc; + } + + if (having_predicate_oper) { + if (*last_oper) { + having_predicate_oper->add_child(std::move(*last_oper)); + } + last_oper = &having_predicate_oper; + } + + if (!select_stmt->order_by().empty()) { + unique_ptr orderby_oper; + rc = create_order_by_plan(select_stmt, orderby_oper); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to create orderby logical plan. rc=%s", strrc(rc)); + return rc; + } + if (orderby_oper) { + if (*last_oper) { + orderby_oper->add_child(std::move(*last_oper)); + } + *last_oper = std::move(orderby_oper); + } + } + + if (select_stmt->limit() != -1) { + unique_ptr limit_oper = std::make_unique(select_stmt->limit()); + if (*last_oper) { + limit_oper->add_child(std::move(*last_oper)); + } + *last_oper = std::move(limit_oper); + } + auto project_oper = make_unique(std::move(select_stmt->query_expressions())); if (*last_oper) { project_oper->add_child(std::move(*last_oper)); @@ -149,95 +199,37 @@ RC LogicalPlanGenerator::create_plan(SelectStmt *select_stmt, unique_ptr &logical_operator) { - RC rc = RC::SUCCESS; - std::vector> cmp_exprs; - const std::vector &filter_units = filter_stmt->filter_units(); - for (const FilterUnit *filter_unit : filter_units) { - const FilterObj &filter_obj_left = filter_unit->left(); - const FilterObj &filter_obj_right = filter_unit->right(); - - unique_ptr left(filter_obj_left.is_attr - ? static_cast(new FieldExpr(filter_obj_left.field)) - : static_cast(new ValueExpr(filter_obj_left.value))); - - unique_ptr right(filter_obj_right.is_attr - ? static_cast(new FieldExpr(filter_obj_right.field)) - : static_cast(new ValueExpr(filter_obj_right.value))); - - if (left->value_type() != right->value_type()) { - auto left_to_right_cost = implicit_cast_cost(left->value_type(), right->value_type()); - auto right_to_left_cost = implicit_cast_cost(right->value_type(), left->value_type()); - if (left_to_right_cost <= right_to_left_cost && left_to_right_cost != INT32_MAX) { - ExprType left_type = left->type(); - auto cast_expr = make_unique(std::move(left), right->value_type()); - if (left_type == ExprType::VALUE) { - Value left_val; - if (OB_FAIL(rc = cast_expr->try_get_value(left_val))) - { - LOG_WARN("failed to get value from left child", strrc(rc)); - return rc; - } - left = make_unique(left_val); - } else { - left = std::move(cast_expr); - } - } else if (right_to_left_cost < left_to_right_cost && right_to_left_cost != INT32_MAX) { - ExprType right_type = right->type(); - auto cast_expr = make_unique(std::move(right), left->value_type()); - if (right_type == ExprType::VALUE) { - Value right_val; - if (OB_FAIL(rc = cast_expr->try_get_value(right_val))) - { - LOG_WARN("failed to get value from right child", strrc(rc)); - return rc; - } - right = make_unique(right_val); - } else { - right = std::move(cast_expr); - } - - } else { - rc = RC::UNSUPPORTED; - LOG_WARN("unsupported cast from %s to %s", attr_type_to_string(left->value_type()), attr_type_to_string(right->value_type())); - return rc; - } - } + RC rc = RC::SUCCESS; - ComparisonExpr *cmp_expr = new ComparisonExpr(filter_unit->comp(), std::move(left), std::move(right)); - cmp_exprs.emplace_back(cmp_expr); + if (filter_stmt == nullptr || filter_stmt->condition_empty()) { + return {}; } - unique_ptr predicate_oper; - if (!cmp_exprs.empty()) { - unique_ptr conjunction_expr(new ConjunctionExpr(ConjunctionExpr::Type::AND, cmp_exprs)); - predicate_oper = unique_ptr(new PredicateLogicalOperator(std::move(conjunction_expr))); + // 递归遍历 condition 检查所有子查询 + rc = ExpressionIterator::condition_iterate_expr(filter_stmt->condition()); + + if (OB_FAIL(rc)) { + return rc; } - logical_operator = std::move(predicate_oper); - return rc; -} + // 构建 ConjunctionExpr 并将其传递给 logical_operator + unique_ptr conjunction_expr( + new ConjunctionExpr(ConjunctionExpr::Type::AND, std::move(filter_stmt->condition()))); + logical_operator = std::make_unique(std::move(conjunction_expr)); -int LogicalPlanGenerator::implicit_cast_cost(AttrType from, AttrType to) -{ - if (from == to) { - return 0; - } - return DataType::type_instance(from)->cast_cost(to); + return rc; } RC LogicalPlanGenerator::create_plan(InsertStmt *insert_stmt, unique_ptr &logical_operator) { - Table *table = insert_stmt->table(); - vector values(insert_stmt->values(), insert_stmt->values() + insert_stmt->value_amount()); - - InsertLogicalOperator *insert_operator = new InsertLogicalOperator(table, values); + auto insert_operator = new InsertLogicalOperator(insert_stmt->table(), insert_stmt->values_list()); logical_operator.reset(insert_operator); return RC::SUCCESS; } RC LogicalPlanGenerator::create_plan(DeleteStmt *delete_stmt, unique_ptr &logical_operator) { - Table *table = delete_stmt->table(); + BaseTable *table = delete_stmt->table(); FilterStmt *filter_stmt = delete_stmt->filter_stmt(); unique_ptr table_get_oper(new TableGetLogicalOperator(table, ReadWriteMode::READ_WRITE)); @@ -261,6 +253,35 @@ RC LogicalPlanGenerator::create_plan(DeleteStmt *delete_stmt, unique_ptr &logical_operator) +{ + auto table = update_stmt->table(); + auto &field_metas = update_stmt->field_metas(); + auto &values = update_stmt->values(); // 支持了多个值 + auto filter_stmt = update_stmt->filter_stmt(); + + auto table_get_oper = std::make_unique(table, ReadWriteMode::READ_WRITE); + + unique_ptr predicate_oper; + + RC rc = create_plan(filter_stmt, predicate_oper); + if (rc != RC::SUCCESS) { + return rc; + } + + auto update_oper = std::make_unique(table, std::move(field_metas), std::move(values)); + + if (predicate_oper) { + predicate_oper->add_child(std::move(table_get_oper)); + update_oper->add_child(std::move(predicate_oper)); + } else { + update_oper->add_child(std::move(table_get_oper)); + } + + logical_operator = std::move(update_oper); + return rc; +} + RC LogicalPlanGenerator::create_plan(ExplainStmt *explain_stmt, unique_ptr &logical_operator) { unique_ptr child_oper; @@ -281,9 +302,15 @@ RC LogicalPlanGenerator::create_plan(ExplainStmt *explain_stmt, unique_ptr &logical_operator) { vector> &group_by_expressions = select_stmt->group_by(); - vector aggregate_expressions; + vector aggregate_expressions; vector> &query_expressions = select_stmt->query_expressions(); - function&)> collector = [&](unique_ptr &expr) -> RC { + + // 还要考虑having里面的 + if (select_stmt->having_filter_stmt() != nullptr) + ExpressionIterator::having_condition_iterate_expr( + select_stmt->having_filter_stmt()->condition(), aggregate_expressions); + + function &)> collector = [&](unique_ptr &expr) -> RC { RC rc = RC::SUCCESS; if (expr->type() == ExprType::AGGREGATION) { expr->set_pos(aggregate_expressions.size() + group_by_expressions.size()); @@ -293,7 +320,7 @@ RC LogicalPlanGenerator::create_group_by_plan(SelectStmt *select_stmt, unique_pt return rc; }; - function&)> bind_group_by_expr = [&](unique_ptr &expr) -> RC { + function &)> bind_group_by_expr = [&](unique_ptr &expr) -> RC { RC rc = RC::SUCCESS; for (size_t i = 0; i < group_by_expressions.size(); i++) { auto &group_by = group_by_expressions[i]; @@ -309,8 +336,8 @@ RC LogicalPlanGenerator::create_group_by_plan(SelectStmt *select_stmt, unique_pt return rc; }; - bool found_unbound_column = false; - function&)> find_unbound_column = [&](unique_ptr &expr) -> RC { + bool found_unbound_column = false; + function &)> find_unbound_column = [&](unique_ptr &expr) -> RC { RC rc = RC::SUCCESS; if (expr->type() == ExprType::AGGREGATION) { // do nothing @@ -318,12 +345,11 @@ RC LogicalPlanGenerator::create_group_by_plan(SelectStmt *select_stmt, unique_pt // do nothing } else if (expr->type() == ExprType::FIELD) { found_unbound_column = true; - }else { + } else { rc = ExpressionIterator::iterate_child_expr(*expr, find_unbound_column); } return rc; }; - for (unique_ptr &expression : query_expressions) { bind_group_by_expr(expression); @@ -350,8 +376,20 @@ RC LogicalPlanGenerator::create_group_by_plan(SelectStmt *select_stmt, unique_pt // 如果只需要聚合，但是没有group by 语句，需要生成一个空的group by 语句 - auto group_by_oper = make_unique(std::move(group_by_expressions), - std::move(aggregate_expressions)); + auto group_by_oper = + make_unique(std::move(group_by_expressions), std::move(aggregate_expressions)); logical_operator = std::move(group_by_oper); return RC::SUCCESS; -} \ No newline at end of file +} + +RC LogicalPlanGenerator::create_order_by_plan(SelectStmt *select_stmt, unique_ptr &logical_operator) +{ + if (select_stmt == nullptr) { + logical_operator = nullptr; + return RC::SUCCESS; + } + + unique_ptr orderby_oper(new OrderByLogicalOperator(std::move(select_stmt->order_by()))); + logical_operator = std::move(orderby_oper); + return RC::SUCCESS; +} diff --git a/src/observer/sql/optimizer/logical_plan_generator.h b/src/observer/sql/optimizer/logical_plan_generator.h index d02cb181..4edb6657 100644 --- a/src/observer/sql/optimizer/logical_plan_generator.h +++ b/src/observer/sql/optimizer/logical_plan_generator.h @@ -25,6 +25,7 @@ class SelectStmt; class FilterStmt; class InsertStmt; class DeleteStmt; +class UpdateStmt; class ExplainStmt; class LogicalOperator; @@ -34,17 +35,16 @@ class LogicalPlanGenerator LogicalPlanGenerator() = default; virtual ~LogicalPlanGenerator() = default; - RC create(Stmt *stmt, std::unique_ptr &logical_operator); + static RC create(Stmt *stmt, std::unique_ptr &logical_operator); private: - RC create_plan(CalcStmt *calc_stmt, std::unique_ptr &logical_operator); - RC create_plan(SelectStmt *select_stmt, std::unique_ptr &logical_operator); - RC create_plan(FilterStmt *filter_stmt, std::unique_ptr &logical_operator); - RC create_plan(InsertStmt *insert_stmt, std::unique_ptr &logical_operator); - RC create_plan(DeleteStmt *delete_stmt, std::unique_ptr &logical_operator); - RC create_plan(ExplainStmt *explain_stmt, std::unique_ptr &logical_operator); - - RC create_group_by_plan(SelectStmt *select_stmt, std::unique_ptr &logical_operator); - - int implicit_cast_cost(AttrType from, AttrType to); -}; \ No newline at end of file + static RC create_plan(CalcStmt *calc_stmt, std::unique_ptr &logical_operator); + static RC create_plan(SelectStmt *select_stmt, std::unique_ptr &logical_operator); + static RC create_plan(FilterStmt *filter_stmt, std::unique_ptr &logical_operator); + static RC create_plan(InsertStmt *insert_stmt, std::unique_ptr &logical_operator); + static RC create_plan(DeleteStmt *delete_stmt, std::unique_ptr &logical_operator); + static RC create_plan(UpdateStmt *update_stmt, std::unique_ptr &logical_operator); + static RC create_plan(ExplainStmt *explain_stmt, std::unique_ptr &logical_operator); + static RC create_group_by_plan(SelectStmt *select_stmt, std::unique_ptr &logical_operator); + static RC create_order_by_plan(SelectStmt *select_stmt, std::unique_ptr &logical_operator); +}; diff --git a/src/observer/sql/optimizer/optimize_stage.cpp b/src/observer/sql/optimizer/optimize_stage.cpp index 902c2ce8..81883f1a 100644 --- a/src/observer/sql/optimizer/optimize_stage.cpp +++ b/src/observer/sql/optimizer/optimize_stage.cpp @@ -77,10 +77,11 @@ RC OptimizeStage::generate_physical_plan( unique_ptr &logical_operator, unique_ptr &physical_operator, Session *session) { RC rc = RC::SUCCESS; - if (session->get_execution_mode() == ExecutionMode::CHUNK_ITERATOR && LogicalOperator::can_generate_vectorized_operator(logical_operator->type())) { + if (session->get_execution_mode() == ExecutionMode::CHUNK_ITERATOR && + LogicalOperator::can_generate_vectorized_operator(logical_operator->type())) { LOG_INFO("use chunk iterator"); session->set_used_chunk_mode(true); - rc = physical_plan_generator_.create_vec(*logical_operator, physical_operator); + rc = physical_plan_generator_.create_vec(*logical_operator, physical_operator); } else { LOG_INFO("use tuple iterator"); session->set_used_chunk_mode(false); diff --git a/src/observer/sql/optimizer/physical_plan_generator.cpp b/src/observer/sql/optimizer/physical_plan_generator.cpp index 9e946c2d..9fc0aa65 100644 --- a/src/observer/sql/optimizer/physical_plan_generator.cpp +++ b/src/observer/sql/optimizer/physical_plan_generator.cpp @@ -26,6 +26,9 @@ See the Mulan PSL v2 for more details. */ #include "sql/operator/expr_vec_physical_operator.h" #include "sql/operator/group_by_vec_physical_operator.h" #include "sql/operator/index_scan_physical_operator.h" +#include "sql/operator/order_by_logical_operator.h" +#include "sql/operator/limit_logical_operator.h" +#include "sql/operator/limit_physical_operator.h" #include "sql/operator/insert_logical_operator.h" #include "sql/operator/insert_physical_operator.h" #include "sql/operator/join_logical_operator.h" @@ -37,12 +40,18 @@ See the Mulan PSL v2 for more details. */ #include "sql/operator/project_vec_physical_operator.h" #include "sql/operator/table_get_logical_operator.h" #include "sql/operator/table_scan_physical_operator.h" +#include "sql/operator/order_by_physical_operator.h" #include "sql/operator/group_by_logical_operator.h" #include "sql/operator/group_by_physical_operator.h" #include "sql/operator/hash_group_by_physical_operator.h" #include "sql/operator/scalar_group_by_physical_operator.h" #include "sql/operator/table_scan_vec_physical_operator.h" #include "sql/optimizer/physical_plan_generator.h" +#include "sql/operator/update_logical_operator.h" +#include "sql/operator/update_physical_operator.h" +#include "sql/operator/view_scan_physical_operator.h" + +#include using namespace std; @@ -75,6 +84,10 @@ RC PhysicalPlanGenerator::create(LogicalOperator &logical_operator, unique_ptr

(logical_operator), oper); } break; + case LogicalOperatorType::UPDATE: { + return create_plan(static_cast(logical_operator), oper); + } break; + case LogicalOperatorType::EXPLAIN: { return create_plan(static_cast(logical_operator), oper); } break; @@ -87,6 +100,14 @@ RC PhysicalPlanGenerator::create(LogicalOperator &logical_operator, unique_ptr

(logical_operator), oper); } break; + case LogicalOperatorType::ORDER_BY: { + return create_plan(static_cast(logical_operator), oper); + } break; + + case LogicalOperatorType::LIMIT: { + return create_plan(static_cast(logical_operator), oper); + } break; + default: { ASSERT(false, "unknown logical operator type"); return RC::INVALID_ARGUMENT; @@ -119,50 +140,70 @@ RC PhysicalPlanGenerator::create_vec(LogicalOperator &logical_operator, unique_p return rc; } - - RC PhysicalPlanGenerator::create_plan(TableGetLogicalOperator &table_get_oper, unique_ptr &oper) { vector> &predicates = table_get_oper.predicates(); // 看看是否有可以用于索引查找的表达式 - Table *table = table_get_oper.table(); + auto base_table = table_get_oper.table(); Index *index = nullptr; ValueExpr *value_expr = nullptr; - for (auto &expr : predicates) { - if (expr->type() == ExprType::COMPARISON) { - auto comparison_expr = static_cast(expr.get()); - // 简单处理，就找等值查询 - if (comparison_expr->comp() != EQUAL_TO) { - continue; - } - - unique_ptr &left_expr = comparison_expr->left(); - unique_ptr &right_expr = comparison_expr->right(); - // 左右比较的一边最少是一个值 - if (left_expr->type() != ExprType::VALUE && right_expr->type() != ExprType::VALUE) { - continue; - } - - FieldExpr *field_expr = nullptr; - if (left_expr->type() == ExprType::FIELD) { - ASSERT(right_expr->type() == ExprType::VALUE, "right expr should be a value expr while left is field expr"); - field_expr = static_cast(left_expr.get()); - value_expr = static_cast(right_expr.get()); - } else if (right_expr->type() == ExprType::FIELD) { - ASSERT(left_expr->type() == ExprType::VALUE, "left expr should be a value expr while right is a field expr"); - field_expr = static_cast(right_expr.get()); - value_expr = static_cast(left_expr.get()); - } - - if (field_expr == nullptr) { - continue; - } + Table *table = nullptr; + + if (base_table->type() == TableType::Table) { + table = dynamic_cast

(base_table); + + // 是向量索引 + if (table_get_oper.is_vector_scan()) { + auto vector_scan_oper = new VectorScanPhysicalOperator(table, + std::move(table_get_oper.table_alias()), + table_get_oper.index(), + table_get_oper.base_vector(), + table_get_oper.limit(), + table_get_oper.read_write_mode()); + + vector_scan_oper->set_predicates(std::move(predicates)); + oper = unique_ptr(vector_scan_oper); + LOG_TRACE("Vector Index scan used on table: {}", table->name()); + return RC::SUCCESS; + } - const Field &field = field_expr->field(); - index = table->find_index_by_field(field.field_name()); - if (nullptr != index) { - break; + for (auto &expr : predicates) { + if (expr->type() == ExprType::COMPARISON) { + auto comparison_expr = dynamic_cast(expr.get()); + + // 简单处理，就找等值查询 + if (comparison_expr->comp() != EQUAL_TO) { + continue; + } + + unique_ptr &left_expr = comparison_expr->left(); + unique_ptr &right_expr = comparison_expr->right(); + // 左右比较的一边最少是一个值 + if (left_expr->type() != ExprType::VALUE && right_expr->type() != ExprType::VALUE) { + continue; + } + + FieldExpr *field_expr = nullptr; + if (left_expr->type() == ExprType::FIELD) { + ASSERT(right_expr->type() == ExprType::VALUE, "right expr should be a value expr while left is field expr"); + field_expr = dynamic_cast(left_expr.get()); + value_expr = dynamic_cast(right_expr.get()); + } else if (right_expr->type() == ExprType::FIELD) { + ASSERT(left_expr->type() == ExprType::VALUE, "left expr should be a value expr while right is a field expr"); + field_expr = dynamic_cast(right_expr.get()); + value_expr = dynamic_cast(left_expr.get()); + } + + if (field_expr == nullptr) { + continue; + } + + const Field &field = field_expr->field(); + index = table->find_index_by_field(field.field_name()); + if (nullptr != index) { + break; + } } } } @@ -172,6 +213,7 @@ RC PhysicalPlanGenerator::create_plan(TableGetLogicalOperator &table_get_oper, u const Value &value = value_expr->get_value(); IndexScanPhysicalOperator *index_scan_oper = new IndexScanPhysicalOperator(table, + std::move(table_get_oper.table_alias()), index, table_get_oper.read_write_mode(), &value, @@ -181,12 +223,24 @@ RC PhysicalPlanGenerator::create_plan(TableGetLogicalOperator &table_get_oper, u index_scan_oper->set_predicates(std::move(predicates)); oper = unique_ptr(index_scan_oper); - LOG_TRACE("use index scan"); + LOG_TRACE("Index scan used on table: {}", table->name()); } else { - auto table_scan_oper = new TableScanPhysicalOperator(table, table_get_oper.read_write_mode()); - table_scan_oper->set_predicates(std::move(predicates)); - oper = unique_ptr(table_scan_oper); - LOG_TRACE("use table scan"); + if (base_table->type() == TableType::Table) { + auto table_scan_oper = std::make_unique( + table, std::move(table_get_oper.table_alias()), table_get_oper.read_write_mode()); + table_scan_oper->set_predicates(std::move(predicates)); + oper = std::move(table_scan_oper); + LOG_TRACE("Table scan used on table: {}", table->name()); + } else if (base_table->type() == TableType::View) { + auto view = dynamic_cast(base_table); + auto table_scan_oper = std::make_unique(view, std::move(table_get_oper.table_alias())); + table_scan_oper->set_predicates(std::move(predicates)); + oper = std::move(table_scan_oper); + LOG_TRACE("View scan used on view: {}", view->name()); + } else { + LOG_ERROR("Unsupported table type: {}", base_table->type()); + return RC::UNKNOWN_TABLE_TYPE; + } } return RC::SUCCESS; @@ -245,9 +299,7 @@ RC PhysicalPlanGenerator::create_plan(ProjectLogicalOperator &project_oper, uniq RC PhysicalPlanGenerator::create_plan(InsertLogicalOperator &insert_oper, unique_ptr &oper) { - Table *table = insert_oper.table(); - vector &values = insert_oper.values(); - InsertPhysicalOperator *insert_phy_oper = new InsertPhysicalOperator(table, std::move(values)); + auto insert_phy_oper = new InsertPhysicalOperator(insert_oper.table(), insert_oper.values_list()); oper.reset(insert_phy_oper); return RC::SUCCESS; } @@ -277,6 +329,32 @@ RC PhysicalPlanGenerator::create_plan(DeleteLogicalOperator &delete_oper, unique return rc; } +RC PhysicalPlanGenerator::create_plan(UpdateLogicalOperator &update_oper, std::unique_ptr &oper) +{ + vector> &child_opers = update_oper.children(); + + unique_ptr child_physical_oper; + + RC rc = RC::SUCCESS; + if (!child_opers.empty()) { + LogicalOperator *child_oper = child_opers.front().get(); + + rc = create(*child_oper, child_physical_oper); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to create physical operator. rc=%s", strrc(rc)); + return rc; + } + } + + oper = std::make_unique( + update_oper.table(), std::move(update_oper.field_metas()), std::move(update_oper.values())); + + if (child_physical_oper) { + oper->add_child(std::move(child_physical_oper)); + } + return rc; +} + RC PhysicalPlanGenerator::create_plan(ExplainLogicalOperator &explain_oper, unique_ptr &oper) { vector> &child_opers = explain_oper.children(); @@ -338,13 +416,13 @@ RC PhysicalPlanGenerator::create_plan(GroupByLogicalOperator &logical_oper, std: { RC rc = RC::SUCCESS; - vector> &group_by_expressions = logical_oper.group_by_expressions(); + vector> &group_by_expressions = logical_oper.group_by_expressions(); unique_ptr group_by_oper; if (group_by_expressions.empty()) { group_by_oper = make_unique(std::move(logical_oper.aggregate_expressions())); } else { - group_by_oper = make_unique(std::move(logical_oper.group_by_expressions()), - std::move(logical_oper.aggregate_expressions())); + group_by_oper = make_unique( + std::move(logical_oper.group_by_expressions()), std::move(logical_oper.aggregate_expressions())); } ASSERT(logical_oper.children().size() == 1, "group by operator should have 1 child"); @@ -363,28 +441,78 @@ RC PhysicalPlanGenerator::create_plan(GroupByLogicalOperator &logical_oper, std: return rc; } +RC PhysicalPlanGenerator::create_plan(OrderByLogicalOperator &logical_oper, std::unique_ptr &oper) +{ + vector> &child_opers = logical_oper.children(); + unique_ptr child_phy_oper; + + RC rc = RC::SUCCESS; + if (!child_opers.empty()) { + LogicalOperator *child_oper = child_opers.front().get(); + rc = create(*child_oper, child_phy_oper); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to create orderby logical operator's child physical operator. rc=%s", strrc(rc)); + return rc; + } + } + + OrderByPhysicalOperator *orderby_operator = new OrderByPhysicalOperator(std::move(logical_oper.order_by())); + if (child_phy_oper) { + orderby_operator->add_child(std::move(child_phy_oper)); + } + + oper = unique_ptr(orderby_operator); + return rc; +} + +RC PhysicalPlanGenerator::create_plan(LimitLogicalOperator &logical_oper, unique_ptr &oper) +{ + vector> &child_opers = logical_oper.children(); + unique_ptr child_phy_oper; + + RC rc = RC::SUCCESS; + if (!child_opers.empty()) { + LogicalOperator *child_oper = child_opers.front().get(); + rc = create(*child_oper, child_phy_oper); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to create orderby logical operator's child physical operator. rc=%s", strrc(rc)); + return rc; + } + } + + LimitPhysicalOperator *limit_oper = new LimitPhysicalOperator(logical_oper.limit()); + limit_oper->add_child(std::move(child_phy_oper)); + + oper = unique_ptr(limit_oper); + return rc; +} + RC PhysicalPlanGenerator::create_vec_plan(TableGetLogicalOperator &table_get_oper, unique_ptr &oper) { vector> &predicates = table_get_oper.predicates(); - Table *table = table_get_oper.table(); - TableScanVecPhysicalOperator *table_scan_oper = new TableScanVecPhysicalOperator(table, table_get_oper.read_write_mode()); - table_scan_oper->set_predicates(std::move(predicates)); - oper = unique_ptr(table_scan_oper); - LOG_TRACE("use vectorized table scan"); + BaseTable *base_table = table_get_oper.table(); + Table *table = nullptr; + if (base_table->type() == TableType::Table) { + table = static_cast
(base_table); + TableScanVecPhysicalOperator *table_scan_oper = + new TableScanVecPhysicalOperator(table, table_get_oper.read_write_mode()); + table_scan_oper->set_predicates(std::move(predicates)); + oper = unique_ptr(table_scan_oper); + LOG_TRACE("use vectorized table scan"); + } return RC::SUCCESS; } RC PhysicalPlanGenerator::create_vec_plan(GroupByLogicalOperator &logical_oper, unique_ptr &oper) { - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; unique_ptr physical_oper = nullptr; if (logical_oper.group_by_expressions().empty()) { physical_oper = make_unique(std::move(logical_oper.aggregate_expressions())); } else { physical_oper = make_unique( - std::move(logical_oper.group_by_expressions()), std::move(logical_oper.aggregate_expressions())); - + std::move(logical_oper.group_by_expressions()), std::move(logical_oper.aggregate_expressions())); } ASSERT(logical_oper.children().size() == 1, "group by operator should have 1 child"); @@ -401,8 +529,6 @@ RC PhysicalPlanGenerator::create_vec_plan(GroupByLogicalOperator &logical_oper, oper = std::move(physical_oper); return rc; - - return RC::SUCCESS; } RC PhysicalPlanGenerator::create_vec_plan(ProjectLogicalOperator &project_oper, unique_ptr &oper) @@ -439,7 +565,6 @@ RC PhysicalPlanGenerator::create_vec_plan(ProjectLogicalOperator &project_oper, return rc; } - RC PhysicalPlanGenerator::create_vec_plan(ExplainLogicalOperator &explain_oper, unique_ptr &oper) { vector> &child_opers = explain_oper.children(); @@ -460,4 +585,4 @@ RC PhysicalPlanGenerator::create_vec_plan(ExplainLogicalOperator &explain_oper, oper = std::move(explain_physical_oper); return rc; -} \ No newline at end of file +} diff --git a/src/observer/sql/optimizer/physical_plan_generator.h b/src/observer/sql/optimizer/physical_plan_generator.h index bac779a0..3aa88326 100644 --- a/src/observer/sql/optimizer/physical_plan_generator.h +++ b/src/observer/sql/optimizer/physical_plan_generator.h @@ -23,10 +23,13 @@ class PredicateLogicalOperator; class ProjectLogicalOperator; class InsertLogicalOperator; class DeleteLogicalOperator; +class UpdateLogicalOperator; class ExplainLogicalOperator; class JoinLogicalOperator; class CalcLogicalOperator; class GroupByLogicalOperator; +class OrderByLogicalOperator; +class LimitLogicalOperator; /** * @brief 物理计划生成器 @@ -40,21 +43,24 @@ class PhysicalPlanGenerator PhysicalPlanGenerator() = default; virtual ~PhysicalPlanGenerator() = default; - RC create(LogicalOperator &logical_operator, std::unique_ptr &oper); - RC create_vec(LogicalOperator &logical_operator, std::unique_ptr &oper); + static RC create(LogicalOperator &logical_operator, std::unique_ptr &oper); + static RC create_vec(LogicalOperator &logical_operator, std::unique_ptr &oper); private: - RC create_plan(TableGetLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_plan(PredicateLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_plan(ProjectLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_plan(InsertLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_plan(DeleteLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_plan(ExplainLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_plan(JoinLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_plan(CalcLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_plan(GroupByLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_vec_plan(ProjectLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_vec_plan(TableGetLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_vec_plan(GroupByLogicalOperator &logical_oper, std::unique_ptr &oper); - RC create_vec_plan(ExplainLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(TableGetLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(PredicateLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(ProjectLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(InsertLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(DeleteLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(UpdateLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(ExplainLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(JoinLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(CalcLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(GroupByLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(OrderByLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_plan(LimitLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_vec_plan(ProjectLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_vec_plan(TableGetLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_vec_plan(GroupByLogicalOperator &logical_oper, std::unique_ptr &oper); + static RC create_vec_plan(ExplainLogicalOperator &logical_oper, std::unique_ptr &oper); }; diff --git a/src/observer/sql/optimizer/predicate_pushdown_rewriter.cpp b/src/observer/sql/optimizer/predicate_pushdown_rewriter.cpp index 342218f7..6d4fd219 100644 --- a/src/observer/sql/optimizer/predicate_pushdown_rewriter.cpp +++ b/src/observer/sql/optimizer/predicate_pushdown_rewriter.cpp @@ -96,8 +96,8 @@ RC PredicatePushdownRewriter::get_exprs_can_pushdown( ConjunctionExpr *conjunction_expr = static_cast(expr.get()); // 或 操作的比较，太复杂，现在不考虑 if (conjunction_expr->conjunction_type() == ConjunctionExpr::Type::OR) { - LOG_WARN("unsupported or operation"); - rc = RC::UNIMPLEMENTED; + // TODO LOG_WARN("unsupported or operation"); + rc = RC::SUCCESS; return rc; } @@ -119,7 +119,7 @@ RC PredicatePushdownRewriter::get_exprs_can_pushdown( } } else if (expr->type() == ExprType::COMPARISON) { // 如果是比较操作，并且比较的左边或右边是表某个列值，那么就下推下去 - auto comparison_expr = static_cast(expr.get()); + auto comparison_expr = static_cast(expr.get()); std::unique_ptr &left_expr = comparison_expr->left(); std::unique_ptr &right_expr = comparison_expr->right(); diff --git a/src/observer/sql/optimizer/rewriter.cpp b/src/observer/sql/optimizer/rewriter.cpp index 8d1cebce..02201f15 100644 --- a/src/observer/sql/optimizer/rewriter.cpp +++ b/src/observer/sql/optimizer/rewriter.cpp @@ -18,12 +18,14 @@ See the Mulan PSL v2 for more details. */ #include "sql/optimizer/expression_rewriter.h" #include "sql/optimizer/predicate_pushdown_rewriter.h" #include "sql/optimizer/predicate_rewrite.h" +#include "sql/optimizer/vector_index_scan_rewrite.h" Rewriter::Rewriter() { rewrite_rules_.emplace_back(new ExpressionRewriter); rewrite_rules_.emplace_back(new PredicateRewriteRule); rewrite_rules_.emplace_back(new PredicatePushdownRewriter); + rewrite_rules_.emplace_back(new VectorIndexScanRewrite); } RC Rewriter::rewrite(std::unique_ptr &oper, bool &change_made) diff --git a/src/observer/sql/optimizer/vector_index_scan_rewrite.cpp b/src/observer/sql/optimizer/vector_index_scan_rewrite.cpp new file mode 100644 index 00000000..7ac0a3c9 --- /dev/null +++ b/src/observer/sql/optimizer/vector_index_scan_rewrite.cpp @@ -0,0 +1,109 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/18 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "sql/operator/logical_operator.h" +#include "sql/expr/tuple.h" +#include "sql/operator/limit_logical_operator.h" +#include "sql/operator/order_by_logical_operator.h" +#include "sql/operator/table_get_logical_operator.h" +#include "sql/optimizer/vector_index_scan_rewrite.h" + +RC VectorIndexScanRewrite::rewrite(std::unique_ptr &oper, bool &change_made) +{ + auto &child_opers = oper->children(); + if (child_opers.size() != 1) { + return RC::SUCCESS; + } + + auto &child_oper = child_opers.front(); + if (child_oper->type() != LogicalOperatorType::LIMIT && child_oper->type() != LogicalOperatorType::ORDER_BY) { + return RC::SUCCESS; + } + + // 有 limit 且 orderby 是向量 重写为向量索引 + if (child_oper->type() == LogicalOperatorType::LIMIT) { + auto &orderby_opers = child_oper->children(); + if (orderby_opers.size() != 1) { + return RC::SUCCESS; + } + + auto &orderby_oper = orderby_opers[0]; + if (orderby_oper->type() != LogicalOperatorType::ORDER_BY) { + return RC::SUCCESS; + } + + auto act_orderby_oper = dynamic_cast(orderby_oper.get()); + auto &orderby_node = act_orderby_oper->order_by(); + if (orderby_node.size() != 1 || !orderby_node[0].is_asc) { + return RC::SUCCESS; + } + + auto &expr = orderby_node[0].expr; + if (expr->type() == ExprType::NORMAL_FUNCTION) { + auto func_expr = dynamic_cast(expr.get()); + if (!func_expr->is_vector_distance_func()) { + return RC::SUCCESS; + } + + auto &table_scan_oper = orderby_oper->children()[0]; + if (table_scan_oper->type() != LogicalOperatorType::TABLE_GET) { + return RC::SUCCESS; + } + + auto table_get_oper = dynamic_cast(table_scan_oper.get()); + BaseTable *base_table = table_get_oper->table(); + if (base_table->type() != TableType::Table) { + return RC::SUCCESS; + } + + auto table = dynamic_cast
(base_table); + + // 找到向量列和常量向量，目前只支持向量列与常量向量 + FieldExpr *field_expr; + RowTuple tuple; + Value value; + + if (func_expr->args()[0]->type() == ExprType::FIELD) { + field_expr = dynamic_cast(func_expr->args()[0].get()); + func_expr->args()[1]->get_value(tuple, value); + } else { + field_expr = dynamic_cast(func_expr->args()[1].get()); + func_expr->args()[0]->get_value(tuple, value); + } + + // 检查该列上是否建立了索引 + auto &field = field_expr->field(); + + if (strcmp(table->name(), field.table_name()) != 0) { + return RC::SUCCESS; + } + + Index *index = table->find_vector_index(func_expr->function_type(), field.field_name()); + if (index == nullptr) { + return RC::SUCCESS; + } + + auto limit_oper = dynamic_cast(child_oper.get()); + + // 设置向量索引必要的参数 + table_get_oper->set_index(index); + table_get_oper->set_base_vector(value.get_vector()); + table_get_oper->set_limit(limit_oper->limit()); + + // 直接变成 proj table_get + oper->children()[0] = std::move(table_scan_oper); + change_made = true; + } + } + + return RC::SUCCESS; +} diff --git a/src/observer/sql/optimizer/vector_index_scan_rewrite.h b/src/observer/sql/optimizer/vector_index_scan_rewrite.h new file mode 100644 index 00000000..df22f307 --- /dev/null +++ b/src/observer/sql/optimizer/vector_index_scan_rewrite.h @@ -0,0 +1,29 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/18 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "sql/optimizer/rewrite_rule.h" + +/** + * @brief 向量索引重写规则 + * @ingroup Rewriter + * @details 识别 orderby limit 重写为向量索引 + */ +class VectorIndexScanRewrite : public RewriteRule +{ +public: + VectorIndexScanRewrite() = default; + virtual ~VectorIndexScanRewrite() = default; + + RC rewrite(std::unique_ptr &oper, bool &change_made) override; +}; diff --git a/src/observer/sql/parser/expression_binder.cpp b/src/observer/sql/parser/expression_binder.cpp index 2daf85cf..e9ae900d 100644 --- a/src/observer/sql/parser/expression_binder.cpp +++ b/src/observer/sql/parser/expression_binder.cpp @@ -22,25 +22,39 @@ See the Mulan PSL v2 for more details. */ using namespace std; using namespace common; -Table *BinderContext::find_table(const char *table_name) const +BaseTable *BinderContext::find_table(const char *table_name) const { - auto pred = [table_name](Table *table) { return 0 == strcasecmp(table_name, table->name()); }; - auto iter = ranges::find_if(query_tables_, pred); - if (iter == query_tables_.end()) { + auto iter = this->tables_->find(table_name); + if (iter == tables_->end()) { return nullptr; } - return *iter; + return iter->second; } //////////////////////////////////////////////////////////////////////////////// -static void wildcard_fields(Table *table, vector> &expressions) +void ExpressionBinder::wildcard_fields( + BaseTable *table, std::string table_alias, vector> &expressions, bool multi_tables) { const TableMeta &table_meta = table->table_meta(); const int field_num = table_meta.field_num(); for (int i = table_meta.sys_field_num(); i < field_num; i++) { Field field(table, table_meta.field(i)); FieldExpr *field_expr = new FieldExpr(field); - field_expr->set_name(field.field_name()); + field_expr->set_table_alias(std::move(table_alias)); + // 这里设置了基类的 name 属性 + if (multi_tables) { + // 多表查询带表名 + auto table_name = field_expr->table_name(); + auto field_name = field_expr->field_name(); + // 创建一个字符数组来存储合并后的字符串 + char result[256]; + // 使用 snprintf 合并字符串 + snprintf(result, sizeof(result), "%s.%s", table_name, field_name); + field_expr->set_name(result); + } else { + // 单表查询不带表名 + field_expr->set_name(field.field_name()); + } expressions.emplace_back(field_expr); } } @@ -60,8 +74,8 @@ RC ExpressionBinder::bind_expression(unique_ptr &expr, vector &expr, vector(expr->type())); return RC::INTERNAL; @@ -109,11 +130,15 @@ RC ExpressionBinder::bind_star_expression( auto star_expr = static_cast(expr.get()); - vector
tables_to_wildcard; + if (!is_blank(star_expr->name()) || !is_blank(star_expr->alias())) { + return RC::INVALID_ALIAS; + } + + vector tables_to_wildcard; const char *table_name = star_expr->table_name(); if (!is_blank(table_name) && 0 != strcmp(table_name, "*")) { - Table *table = context_.find_table(table_name); + BaseTable *table = context_.find_table(table_name); if (nullptr == table) { LOG_INFO("no such table in from list: %s", table_name); return RC::SCHEMA_TABLE_NOT_EXIST; @@ -121,12 +146,24 @@ RC ExpressionBinder::bind_star_expression( tables_to_wildcard.push_back(table); } else { - const vector
&all_tables = context_.query_tables(); + const vector &all_tables = context_.query_tables(); tables_to_wildcard.insert(tables_to_wildcard.end(), all_tables.begin(), all_tables.end()); } - for (Table *table : tables_to_wildcard) { - wildcard_fields(table, bound_expressions); + // select t2/table_alias_2.* from table_alias_1 t1, table_alias_2 t2 where t1.id < t2.id; 非自交可能指定也可能没 + // select t2.* from table_alias_1 t1, table_alias_1 t2 where t1.id < t2.id; 自交的话必然指定了别名 + if (tables_to_wildcard.size() == 1) { + // 看看能不能找到对应的表名，能的话是第一种情况 + for (int i = 0; i < context_.query_tables().size(); ++i) { + if (strcmp(context_.query_tables()[i]->name(), table_name) == 0) { + table_name = context_.alias()[i].c_str(); + } + } + wildcard_fields(tables_to_wildcard[0], table_name, bound_expressions, multi_tables_); + } else { + for (int i = 0; i < tables_to_wildcard.size(); ++i) { + wildcard_fields(tables_to_wildcard[i], context_.alias()[i], bound_expressions, multi_tables_); + } } return RC::SUCCESS; @@ -144,14 +181,9 @@ RC ExpressionBinder::bind_unbound_field_expression( const char *table_name = unbound_field_expr->table_name(); const char *field_name = unbound_field_expr->field_name(); - Table *table = nullptr; + BaseTable *table = nullptr; if (is_blank(table_name)) { - if (context_.query_tables().size() != 1) { - LOG_INFO("cannot determine table for field: %s", field_name); - return RC::SCHEMA_TABLE_NOT_EXIST; - } - - table = context_.query_tables()[0]; + table = context_.default_table(); } else { table = context_.find_table(table_name); if (nullptr == table) { @@ -160,8 +192,16 @@ RC ExpressionBinder::bind_unbound_field_expression( } } + std::string table_alias; + if (context_.has_tables_alias()) { + // 多表自交要投影某些列或者有谓词条件的，一定是通过别名进行查询 + if (strcmp(table->name(), table_name) != 0) { + table_alias = table_name; + } + } + if (0 == strcmp(field_name, "*")) { - wildcard_fields(table, bound_expressions); + wildcard_fields(table, table_alias, bound_expressions, multi_tables_); } else { const FieldMeta *field_meta = table->table_meta().field(field_name); if (nullptr == field_meta) { @@ -171,7 +211,20 @@ RC ExpressionBinder::bind_unbound_field_expression( Field field(table, field_meta); FieldExpr *field_expr = new FieldExpr(field); - field_expr->set_name(field_name); + field_expr->set_table_alias(table_alias); + // 这里设置了基类的 name 属性 + if (!is_blank(unbound_field_expr->alias())) { + field_expr->set_name(unbound_field_expr->alias()); + } else if (multi_tables_) { + // 创建一个字符数组来存储合并后的字符串 + char result[256]; + // 使用 snprintf 合并字符串 + snprintf(result, sizeof(result), "%s.%s", table_name, field_name); + field_expr->set_name(result); + } else { + // 单表查询不带表名 + field_expr->set_name(field_name); + } bound_expressions.emplace_back(field_expr); } @@ -181,6 +234,18 @@ RC ExpressionBinder::bind_unbound_field_expression( RC ExpressionBinder::bind_field_expression( unique_ptr &field_expr, vector> &bound_expressions) { + auto field = static_cast(field_expr.get()); + // 这里设置了基类的 name 属性 + if (multi_tables_) { + // 多表查询带表名 + auto table_name = field->table_name(); + auto field_name = field->field_name(); + // 创建一个字符数组来存储合并后的字符串 + char result[256]; + // 使用 snprintf 合并字符串 + snprintf(result, sizeof(result), "%s.%s", table_name, field_name); + field_expr->set_name(result); + } bound_expressions.emplace_back(std::move(field_expr)); return RC::SUCCESS; } @@ -227,45 +292,44 @@ RC ExpressionBinder::bind_cast_expression( RC ExpressionBinder::bind_comparison_expression( unique_ptr &expr, vector> &bound_expressions) { + RC rc = RC::SUCCESS; if (nullptr == expr) { - return RC::SUCCESS; + return rc; } - auto comparison_expr = static_cast(expr.get()); + auto comparison_expr = dynamic_cast(expr.get()); vector> child_bound_expressions; unique_ptr &left_expr = comparison_expr->left(); unique_ptr &right_expr = comparison_expr->right(); - RC rc = bind_expression(left_expr, child_bound_expressions); - if (rc != RC::SUCCESS) { - return rc; - } + if (nullptr != left_expr) { + rc = bind_expression(left_expr, child_bound_expressions); + if (rc != RC::SUCCESS) { + return rc; + } - if (child_bound_expressions.size() != 1) { - LOG_WARN("invalid left children number of comparison expression: %d", child_bound_expressions.size()); - return RC::INVALID_ARGUMENT; - } + if (child_bound_expressions.size() != 1) { + LOG_WARN("invalid left children number of comparison expression: %d", child_bound_expressions.size()); + return RC::INVALID_ARGUMENT; + } - unique_ptr &left = child_bound_expressions[0]; - if (left.get() != left_expr.get()) { - left_expr.reset(left.release()); + unique_ptr &left = child_bound_expressions[0]; + if (left.get() != left_expr.get()) { + left_expr = std::move(left); + } } - child_bound_expressions.clear(); rc = bind_expression(right_expr, child_bound_expressions); if (rc != RC::SUCCESS) { return rc; } - if (child_bound_expressions.size() != 1) { - LOG_WARN("invalid right children number of comparison expression: %d", child_bound_expressions.size()); - return RC::INVALID_ARGUMENT; - } - - unique_ptr &right = child_bound_expressions[0]; - if (right.get() != right_expr.get()) { - right_expr.reset(right.release()); + if (child_bound_expressions.size() == 1) { + unique_ptr &right = child_bound_expressions[0]; + if (right.get() != right_expr.get()) { + right_expr = std::move(right); + } } bound_expressions.emplace_back(std::move(expr)); @@ -299,7 +363,7 @@ RC ExpressionBinder::bind_conjunction_expression( unique_ptr &child = child_bound_expressions[0]; if (child.get() != child_expr.get()) { - child_expr.reset(child.release()); + child_expr = std::move(child); } } @@ -330,10 +394,12 @@ RC ExpressionBinder::bind_arithmetic_expression( LOG_WARN("invalid left children number of comparison expression: %d", child_bound_expressions.size()); return RC::INVALID_ARGUMENT; } - + if (!is_blank(expr->alias())) { + expr->set_name(expr->alias()); + } unique_ptr &left = child_bound_expressions[0]; if (left.get() != left_expr.get()) { - left_expr.reset(left.release()); + left_expr = std::move(left); } child_bound_expressions.clear(); @@ -342,21 +408,18 @@ RC ExpressionBinder::bind_arithmetic_expression( return rc; } - if (child_bound_expressions.size() != 1) { - LOG_WARN("invalid right children number of comparison expression: %d", child_bound_expressions.size()); - return RC::INVALID_ARGUMENT; - } - - unique_ptr &right = child_bound_expressions[0]; - if (right.get() != right_expr.get()) { - right_expr.reset(right.release()); + if (child_bound_expressions.size() == 1) { + unique_ptr &right = child_bound_expressions[0]; + if (right.get() != right_expr.get()) { + right_expr = std::move(right); + } } bound_expressions.emplace_back(std::move(expr)); return RC::SUCCESS; } -RC check_aggregate_expression(AggregateExpr &expression) +RC check_aggregate_expression(AggregateFunctionExpr &expression) { // 必须有一个子表达式 Expression *child_expression = expression.child().get(); @@ -366,11 +429,11 @@ RC check_aggregate_expression(AggregateExpr &expression) } // 校验数据类型与聚合类型是否匹配 - AggregateExpr::Type aggregate_type = expression.aggregate_type(); - AttrType child_value_type = child_expression->value_type(); + AggregateFunctionType aggregate_type = expression.aggregate_type(); + AttrType child_value_type = child_expression->value_type(); switch (aggregate_type) { - case AggregateExpr::Type::SUM: - case AggregateExpr::Type::AVG: { + case AggregateFunctionType::SUM: + case AggregateFunctionType::AVG: { // 仅支持数值类型 if (child_value_type != AttrType::INTS && child_value_type != AttrType::FLOATS) { LOG_WARN("invalid child value type for aggregate expression: %d", static_cast(child_value_type)); @@ -378,15 +441,16 @@ RC check_aggregate_expression(AggregateExpr &expression) } } break; - case AggregateExpr::Type::COUNT: - case AggregateExpr::Type::MAX: - case AggregateExpr::Type::MIN: { + case AggregateFunctionType::COUNT: + + case AggregateFunctionType::MAX: + case AggregateFunctionType::MIN: { // 任何类型都支持 } break; } // 子表达式中不能再包含聚合表达式 - function&)> check_aggregate_expr = [&](unique_ptr &expr) -> RC { + function &)> check_aggregate_expr = [&](unique_ptr &expr) -> RC { RC rc = RC::SUCCESS; if (expr->type() == ExprType::AGGREGATION) { LOG_WARN("aggregate expression cannot be nested"); @@ -401,51 +465,116 @@ RC check_aggregate_expression(AggregateExpr &expression) return rc; } -RC ExpressionBinder::bind_aggregate_expression( +RC ExpressionBinder::bind_function_expression( unique_ptr &expr, vector> &bound_expressions) { if (nullptr == expr) { return RC::SUCCESS; } - auto unbound_aggregate_expr = static_cast(expr.get()); - const char *aggregate_name = unbound_aggregate_expr->aggregate_name(); - AggregateExpr::Type aggregate_type; - RC rc = AggregateExpr::type_from_string(aggregate_name, aggregate_type); - if (OB_FAIL(rc)) { - LOG_WARN("invalid aggregate name: %s", aggregate_name); - return rc; - } + auto unbound_function_expr = static_cast(expr.get()); + const char *function_name = unbound_function_expr->function_name(); + AggregateFunctionType aggregate_type; + RC rc = AggregateFunctionExpr::type_from_string(function_name, aggregate_type); + if (OB_SUCC(rc)) { + if (unbound_function_expr->args().size() != 1) { + return RC::INVALID_ARGUMENT; + } + unique_ptr &child_expr = unbound_function_expr->args().front(); + vector> child_bound_expressions; + + if (child_expr->type() == ExprType::STAR && aggregate_type == AggregateFunctionType::COUNT) { + ValueExpr *value_expr = new ValueExpr(Value(1)); + child_expr.reset(value_expr); + // count(*) 输出星号 + child_expr->set_name("*"); + unbound_function_expr->set_name(unbound_function_expr->to_string()); + } else { + rc = bind_expression(child_expr, child_bound_expressions); + if (OB_FAIL(rc)) { + return rc; + } - unique_ptr &child_expr = unbound_aggregate_expr->child(); - vector> child_bound_expressions; + if (child_bound_expressions.size() != 1) { + LOG_WARN("invalid children number of aggregate expression: %d", child_bound_expressions.size()); + return RC::INVALID_ARGUMENT; + } - if (child_expr->type() == ExprType::STAR && aggregate_type == AggregateExpr::Type::COUNT) { - ValueExpr *value_expr = new ValueExpr(Value(1)); - child_expr.reset(value_expr); - } else { - rc = bind_expression(child_expr, child_bound_expressions); - if (OB_FAIL(rc)) { - return rc; + if (child_bound_expressions[0].get() != child_expr.get()) { + child_expr.reset(child_bound_expressions[0].release()); + } } - if (child_bound_expressions.size() != 1) { - LOG_WARN("invalid children number of aggregate expression: %d", child_bound_expressions.size()); - return RC::INVALID_ARGUMENT; + auto aggregate_expr = make_unique(aggregate_type, std::move(child_expr)); + + // set name 阶段 + aggregate_expr->set_name(unbound_function_expr->name()); + aggregate_expr->set_alias(unbound_function_expr->alias()); + rc = check_aggregate_expression(*aggregate_expr); + if (OB_FAIL(rc)) { + return rc; } + bound_expressions.emplace_back(std::move(aggregate_expr)); + return RC::SUCCESS; + } - if (child_bound_expressions[0].get() != child_expr.get()) { - child_expr.reset(child_bound_expressions[0].release()); + NormalFunctionType func_type; + rc = NormalFunctionExpr::type_from_string(function_name, func_type); + if (OB_SUCC(rc)) { + vector> child_bound_expressions; + for (auto &child_expr : unbound_function_expr->args()) { + rc = bind_expression(child_expr, child_bound_expressions); + if (OB_FAIL(rc)) { + return rc; + } } + unbound_function_expr->set_args(std::move(child_bound_expressions)); + + string name = unbound_function_expr->name(); + auto func_expr = make_unique( + func_type, unbound_function_expr->function_name(), std::move(unbound_function_expr->args())); + func_expr->set_name(name); + func_expr->set_alias(unbound_function_expr->alias()); + bound_expressions.emplace_back(std::move(func_expr)); + return RC::SUCCESS; } - auto aggregate_expr = make_unique(aggregate_type, std::move(child_expr)); - aggregate_expr->set_name(unbound_aggregate_expr->name()); - rc = check_aggregate_expression(*aggregate_expr); + return RC::UNKNOWN_FUNCTION; +} + +RC ExpressionBinder::bind_subquery_expression( + std::unique_ptr &expr, std::vector> &bound_expressions) +{ + RC rc = RC::SUCCESS; + auto subquery_expr = dynamic_cast(expr.get()); + + rc = subquery_expr->generate_select_stmt(context_.get_db(), context_.table_map()); + if (OB_FAIL(rc)) { + return rc; + } + rc = subquery_expr->generate_logical_oper(); if (OB_FAIL(rc)) { return rc; } + rc = subquery_expr->generate_physical_oper(); - bound_expressions.emplace_back(std::move(aggregate_expr)); - return RC::SUCCESS; + bound_expressions.emplace_back(std::move(expr)); + return rc; +} + +RC ExpressionBinder::bind_exprlist_expression( + std::unique_ptr &expr, std::vector> &bound_expressions) +{ + RC rc = RC::SUCCESS; + auto list_expr = dynamic_cast(expr.get()); + vector> child_bound_expressions; + for (auto &child_expr : list_expr->get_list()) { + if (child_expr->type() != ExprType::VALUE) { + LOG_WARN("invalid children type of LIST expression: %d", child_bound_expressions.size()); + return RC::INVALID_ARGUMENT; + } + } + + bound_expressions.emplace_back(std::move(expr)); + return rc; } diff --git a/src/observer/sql/parser/expression_binder.h b/src/observer/sql/parser/expression_binder.h index d52f049c..7b856536 100644 --- a/src/observer/sql/parser/expression_binder.h +++ b/src/observer/sql/parser/expression_binder.h @@ -24,14 +24,37 @@ class BinderContext BinderContext() = default; virtual ~BinderContext() = default; - void add_table(Table *table) { query_tables_.push_back(table); } + void add_table(BaseTable *table) { query_tables_.emplace_back(table); } + void add_db(Db *db) { db_ = db; } - Table *find_table(const char *table_name) const; + void set_tables(std::unordered_map *tables) { tables_ = tables; } - const std::vector
&query_tables() const { return query_tables_; } + Db *get_db() const { return db_; } + + BaseTable *find_table(const char *table_name) const; + + const std::vector &query_tables() const { return query_tables_; } + std::unordered_map &table_map() { return *tables_; } + + // 对于 update delete 目前还不支持给表取别名，所以是空的 + bool has_tables_alias() { return !tables_alias_.empty(); } + + const std::vector &alias() { return tables_alias_; } + + void set_alias(std::vector alias) { tables_alias_ = std::move(alias); } + +private: + Db *db_; + +public: + [[nodiscard]] BaseTable *default_table() const { return default_table_; } + void set_default_table(BaseTable *default_table) { default_table_ = default_table; } private: - std::vector
query_tables_; + BaseTable *default_table_; + std::vector query_tables_; + std::vector tables_alias_; + std::unordered_map *tables_; }; /** @@ -41,7 +64,8 @@ class BinderContext class ExpressionBinder { public: - ExpressionBinder(BinderContext &context) : context_(context) {} + ExpressionBinder(BinderContext &context) : context_(context) { multi_tables_ = context_.query_tables().size() > 1; } + virtual ~ExpressionBinder() = default; RC bind_expression(std::unique_ptr &expr, std::vector> &bound_expressions); @@ -63,9 +87,17 @@ class ExpressionBinder std::unique_ptr &conjunction_expr, std::vector> &bound_expressions); RC bind_arithmetic_expression( std::unique_ptr &arithmetic_expr, std::vector> &bound_expressions); - RC bind_aggregate_expression( + RC bind_function_expression( std::unique_ptr &aggregate_expr, std::vector> &bound_expressions); + RC bind_subquery_expression( + std::unique_ptr &expr, std::vector> &bound_expressions); + RC bind_exprlist_expression( + std::unique_ptr &expr, std::vector> &bound_expressions); + + void wildcard_fields(BaseTable *table, std::string table_alias, vector> &expressions, + bool multi_tables = false); private: + bool multi_tables_; BinderContext &context_; -}; \ No newline at end of file +}; diff --git a/src/observer/sql/parser/gen_parser.sh b/src/observer/sql/parser/gen_parser.sh index 771fc772..ee463786 100755 --- a/src/observer/sql/parser/gen_parser.sh +++ b/src/observer/sql/parser/gen_parser.sh @@ -1,3 +1,3 @@ #!/bin/bash flex --outfile lex_sql.cpp --header-file=lex_sql.h lex_sql.l -`which bison` -d --output yacc_sql.cpp yacc_sql.y +bison -d --output yacc_sql.cpp yacc_sql.y diff --git a/src/observer/sql/parser/lex_sql.cpp b/src/observer/sql/parser/lex_sql.cpp index f27c1b94..22646f24 100644 --- a/src/observer/sql/parser/lex_sql.cpp +++ b/src/observer/sql/parser/lex_sql.cpp @@ -1,5 +1,4 @@ -#line 2 "lex_sql.cpp" -#line 2 "lex_sql.l" +#line 1 "lex_sql.cpp" /* 这里的代码会被复制到lex_sql.cpp的最开始位置 定义yy_size_t的原因是因为flex生成的代码，会使用yy_size_t与其他类型的数字 @@ -9,37 +8,56 @@ typedef int yy_size_t; /* 参考生成的lex_sql.cpp代码，这个宏定义会放在每次运行yylex()最开始的地方 */ -#define YY_USER_INIT \ - yycolumn = 0; +#define YY_USER_INIT yycolumn = 0; /* 参考生成的lex_sql.cpp代码，这个宏定义会放在解析一个token之后，也可以在网上找到大量的参考资料 */ /* 我们在这里设置当前解析的token的位置信息，这样在yacc中就可以使用这些信息了 */ -#define YY_USER_ACTION \ -do { \ - yylloc->first_line = yylloc->last_line = yylineno; \ - yylloc->first_column = yycolumn; \ - yylloc->last_column = yylloc->first_column + yyleng - 1; \ - yycolumn += yyleng; \ -} \ -while (0); +#define YY_USER_ACTION \ + do { \ + yylloc->first_line = yylloc->last_line = yylineno; \ + yylloc->first_column = yycolumn; \ + yylloc->last_column = yylloc->first_column + yyleng - 1; \ + yycolumn += yyleng; \ + } while (0); +#line 25 "lex_sql.cpp" - - -#line 30 "lex_sql.cpp" - -#define YY_INT_ALIGNED short int +#define YY_INT_ALIGNED short int /* A lexical scanner generated by flex */ #define FLEX_SCANNER #define YY_FLEX_MAJOR_VERSION 2 -#define YY_FLEX_MINOR_VERSION 5 -#define YY_FLEX_SUBMINOR_VERSION 35 +#define YY_FLEX_MINOR_VERSION 6 +#define YY_FLEX_SUBMINOR_VERSION 4 #if YY_FLEX_SUBMINOR_VERSION > 0 #define FLEX_BETA #endif +#ifdef yyget_lval +#define yyget_lval_ALREADY_DEFINED +#else +#define yyget_lval yyget_lval +#endif + +#ifdef yyset_lval +#define yyset_lval_ALREADY_DEFINED +#else +#define yyset_lval yyset_lval +#endif + +#ifdef yyget_lloc +#define yyget_lloc_ALREADY_DEFINED +#else +#define yyget_lloc yyget_lloc +#endif + +#ifdef yyset_lloc +#define yyset_lloc_ALREADY_DEFINED +#else +#define yyset_lloc yyset_lloc +#endif + /* First, we deal with platform-specific or compiler-specific issues. */ /* begin standard C headers. */ @@ -57,97 +75,91 @@ while (0); /* C99 systems have . Non-C99 systems may or may not. */ -#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L +#if defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 says to define __STDC_LIMIT_MACROS before including stdint.h, - * if you want the limit (max/min) macros for int types. + * if you want the limit (max/min) macros for int types. */ #ifndef __STDC_LIMIT_MACROS #define __STDC_LIMIT_MACROS 1 #endif #include -typedef int8_t flex_int8_t; -typedef uint8_t flex_uint8_t; -typedef int16_t flex_int16_t; +typedef int8_t flex_int8_t; +typedef uint8_t flex_uint8_t; +typedef int16_t flex_int16_t; typedef uint16_t flex_uint16_t; -typedef int32_t flex_int32_t; +typedef int32_t flex_int32_t; typedef uint32_t flex_uint32_t; +typedef uint64_t flex_uint64_t; #else -typedef signed char flex_int8_t; -typedef short int flex_int16_t; -typedef int flex_int32_t; -typedef unsigned char flex_uint8_t; +typedef signed char flex_int8_t; +typedef short int flex_int16_t; +typedef int flex_int32_t; +typedef unsigned char flex_uint8_t; typedef unsigned short int flex_uint16_t; -typedef unsigned int flex_uint32_t; -#endif /* ! C99 */ +typedef unsigned int flex_uint32_t; /* Limits of integral types. */ #ifndef INT8_MIN -#define INT8_MIN (-128) +#define INT8_MIN (-128) #endif #ifndef INT16_MIN -#define INT16_MIN (-32767-1) +#define INT16_MIN (-32767 - 1) #endif #ifndef INT32_MIN -#define INT32_MIN (-2147483647-1) +#define INT32_MIN (-2147483647 - 1) #endif #ifndef INT8_MAX -#define INT8_MAX (127) +#define INT8_MAX (127) #endif #ifndef INT16_MAX -#define INT16_MAX (32767) +#define INT16_MAX (32767) #endif #ifndef INT32_MAX -#define INT32_MAX (2147483647) +#define INT32_MAX (2147483647) #endif #ifndef UINT8_MAX -#define UINT8_MAX (255U) +#define UINT8_MAX (255U) #endif #ifndef UINT16_MAX -#define UINT16_MAX (65535U) +#define UINT16_MAX (65535U) #endif #ifndef UINT32_MAX -#define UINT32_MAX (4294967295U) +#define UINT32_MAX (4294967295U) #endif -#endif /* ! FLEXINT_H */ - -#ifdef __cplusplus - -/* The "const" storage-class-modifier is valid. */ -#define YY_USE_CONST - -#else /* ! __cplusplus */ +#ifndef SIZE_MAX +#define SIZE_MAX (~(size_t)0) +#endif -/* C99 requires __STDC__ to be defined as 1. */ -#if defined (__STDC__) +#endif /* ! C99 */ -#define YY_USE_CONST +#endif /* ! FLEXINT_H */ -#endif /* defined (__STDC__) */ -#endif /* ! __cplusplus */ +/* begin standard C++ headers. */ -#ifdef YY_USE_CONST +/* TODO: this is always defined, so inline it */ #define yyconst const + +#if defined(__GNUC__) && __GNUC__ >= 3 +#define yynoreturn __attribute__((__noreturn__)) #else -#define yyconst +#define yynoreturn #endif /* Returned upon end-of-file. */ #define YY_NULL 0 -/* Promotes a possibly negative, possibly signed char to an unsigned - * integer for use as an array index. If the signed char is negative, - * we want to instead treat it as an 8-bit unsigned char, hence the - * double cast. +/* Promotes a possibly negative, possibly signed char to an + * integer in range [0..255] for use as an array index. */ -#define YY_SC_TO_UI(c) ((unsigned int) (unsigned char) c) +#define YY_SC_TO_UI(c) ((YY_CHAR)(c)) /* An opaque pointer. */ #ifndef YY_TYPEDEF_YY_SCANNER_T #define YY_TYPEDEF_YY_SCANNER_T -typedef void* yyscan_t; +typedef void *yyscan_t; #endif /* For convenience, these vars (plus the bison vars far below) @@ -166,126 +178,127 @@ typedef void* yyscan_t; * definition of BEGIN. */ #define BEGIN yyg->yy_start = 1 + 2 * - /* Translate the current start state into a value that can be later handed * to BEGIN to return to the state. The YYSTATE alias is for lex * compatibility. */ #define YY_START ((yyg->yy_start - 1) / 2) #define YYSTATE YY_START - /* Action number for EOF rule of a given start state. */ #define YY_STATE_EOF(state) (YY_END_OF_BUFFER + state + 1) - /* Special action meaning "start processing a new file". */ -#define YY_NEW_FILE yyrestart(yyin ,yyscanner ) - +#define YY_NEW_FILE yyrestart(yyin, yyscanner) #define YY_END_OF_BUFFER_CHAR 0 /* Size of default input buffer. */ #ifndef YY_BUF_SIZE +#ifdef __ia64__ +/* On IA-64, the buffer size is 16k, not 8k. + * Moreover, YY_BUF_SIZE is 2*YY_READ_BUF_SIZE in the general case. + * Ditto for the __ia64__ case accordingly. + */ +#define YY_BUF_SIZE 32768 +#else #define YY_BUF_SIZE 16384 +#endif /* __ia64__ */ #endif /* The state buf must be large enough to hold one state per character in the main buffer. */ -#define YY_STATE_BUF_SIZE ((YY_BUF_SIZE + 2) * sizeof(yy_state_type)) +#define YY_STATE_BUF_SIZE ((YY_BUF_SIZE + 2) * sizeof(yy_state_type)) #ifndef YY_TYPEDEF_YY_BUFFER_STATE #define YY_TYPEDEF_YY_BUFFER_STATE typedef struct yy_buffer_state *YY_BUFFER_STATE; #endif +#ifndef YY_TYPEDEF_YY_SIZE_T +#define YY_TYPEDEF_YY_SIZE_T +typedef size_t yy_size_t; +#endif + #define EOB_ACT_CONTINUE_SCAN 0 #define EOB_ACT_END_OF_FILE 1 #define EOB_ACT_LAST_MATCH 2 - #define YY_LESS_LINENO(n) - -/* Return all but the first "n" matched characters back to the input stream. */ -#define yyless(n) \ - do \ - { \ - /* Undo effects of setting up yytext. */ \ - int yyless_macro_arg = (n); \ - YY_LESS_LINENO(yyless_macro_arg);\ - *yy_cp = yyg->yy_hold_char; \ - YY_RESTORE_YY_MORE_OFFSET \ - yyg->yy_c_buf_p = yy_cp = yy_bp + yyless_macro_arg - YY_MORE_ADJ; \ - YY_DO_BEFORE_ACTION; /* set up yytext again */ \ - } \ - while ( 0 ) - -#define unput(c) yyunput( c, yyg->yytext_ptr , yyscanner ) +#define YY_LESS_LINENO(n) +#define YY_LINENO_REWIND_TO(ptr) -#ifndef YY_TYPEDEF_YY_SIZE_T -#define YY_TYPEDEF_YY_SIZE_T -typedef size_t yy_size_t; -#endif +/* Return all but the first "n" matched characters back to the input stream. */ +#define yyless(n) \ + do { \ + /* Undo effects of setting up yytext. */ \ + int yyless_macro_arg = (n); \ + YY_LESS_LINENO(yyless_macro_arg); \ + *yy_cp = yyg->yy_hold_char; \ + YY_RESTORE_YY_MORE_OFFSET \ + yyg->yy_c_buf_p = yy_cp = yy_bp + yyless_macro_arg - YY_MORE_ADJ; \ + YY_DO_BEFORE_ACTION; /* set up yytext again */ \ + } while (0) +#define unput(c) yyunput(c, yyg->yytext_ptr, yyscanner) #ifndef YY_STRUCT_YY_BUFFER_STATE #define YY_STRUCT_YY_BUFFER_STATE struct yy_buffer_state - { - FILE *yy_input_file; - - char *yy_ch_buf; /* input buffer */ - char *yy_buf_pos; /* current position in input buffer */ - - /* Size of input buffer in bytes, not including room for EOB - * characters. - */ - yy_size_t yy_buf_size; - - /* Number of characters read into yy_ch_buf, not including EOB - * characters. - */ - int yy_n_chars; - - /* Whether we "own" the buffer - i.e., we know we created it, - * and can realloc() it to grow it, and should free() it to - * delete it. - */ - int yy_is_our_buffer; - - /* Whether this is an "interactive" input source; if so, and - * if we're using stdio for input, then we want to use getc() - * instead of fread(), to make sure we stop fetching input after - * each newline. - */ - int yy_is_interactive; - - /* Whether we're considered to be at the beginning of a line. - * If so, '^' rules will be active on the next match, otherwise - * not. - */ - int yy_at_bol; - - int yy_bs_lineno; /**< The line count. */ - int yy_bs_column; /**< The column count. */ - - /* Whether to try to fill the input buffer when we reach the - * end of it. - */ - int yy_fill_buffer; - - int yy_buffer_status; +{ + FILE *yy_input_file; + + char *yy_ch_buf; /* input buffer */ + char *yy_buf_pos; /* current position in input buffer */ + + /* Size of input buffer in bytes, not including room for EOB + * characters. + */ + int yy_buf_size; + + /* Number of characters read into yy_ch_buf, not including EOB + * characters. + */ + yy_size_t yy_n_chars; + + /* Whether we "own" the buffer - i.e., we know we created it, + * and can realloc() it to grow it, and should free() it to + * delete it. + */ + int yy_is_our_buffer; + + /* Whether this is an "interactive" input source; if so, and + * if we're using stdio for input, then we want to use getc() + * instead of fread(), to make sure we stop fetching input after + * each newline. + */ + int yy_is_interactive; + + /* Whether we're considered to be at the beginning of a line. + * If so, '^' rules will be active on the next match, otherwise + * not. + */ + int yy_at_bol; + + int yy_bs_lineno; /**< The line count. */ + int yy_bs_column; /**< The column count. */ + + /* Whether to try to fill the input buffer when we reach the + * end of it. + */ + int yy_fill_buffer; + + int yy_buffer_status; #define YY_BUFFER_NEW 0 #define YY_BUFFER_NORMAL 1 - /* When an EOF's been seen but there's still some text to process - * then we mark the buffer as YY_EOF_PENDING, to indicate that we - * shouldn't try reading from the input source any more. We might - * still have a bunch of tokens to match, though, because of - * possible backing-up. - * - * When we actually see the EOF, we change the status to "new" - * (via yyrestart()), so that the user can continue scanning by - * just pointing yyin at a new input file. - */ + /* When an EOF's been seen but there's still some text to process + * then we mark the buffer as YY_EOF_PENDING, to indicate that we + * shouldn't try reading from the input source any more. We might + * still have a bunch of tokens to match, though, because of + * possible backing-up. + * + * When we actually see the EOF, we change the status to "new" + * (via yyrestart()), so that the user can continue scanning by + * just pointing yyin at a new input file. + */ #define YY_BUFFER_EOF_PENDING 2 - - }; +}; #endif /* !YY_STRUCT_YY_BUFFER_STATE */ /* We provide macros for accessing buffer states in case in the @@ -294,346 +307,2897 @@ struct yy_buffer_state * * Returns the top of the stack, or NULL. */ -#define YY_CURRENT_BUFFER ( yyg->yy_buffer_stack \ - ? yyg->yy_buffer_stack[yyg->yy_buffer_stack_top] \ - : NULL) - +#define YY_CURRENT_BUFFER (yyg->yy_buffer_stack ? yyg->yy_buffer_stack[yyg->yy_buffer_stack_top] : NULL) /* Same as previous macro, but useful when we know that the buffer stack is not * NULL or when we need an lvalue. For internal use only. */ #define YY_CURRENT_BUFFER_LVALUE yyg->yy_buffer_stack[yyg->yy_buffer_stack_top] -void yyrestart (FILE *input_file ,yyscan_t yyscanner ); -void yy_switch_to_buffer (YY_BUFFER_STATE new_buffer ,yyscan_t yyscanner ); -YY_BUFFER_STATE yy_create_buffer (FILE *file,int size ,yyscan_t yyscanner ); -void yy_delete_buffer (YY_BUFFER_STATE b ,yyscan_t yyscanner ); -void yy_flush_buffer (YY_BUFFER_STATE b ,yyscan_t yyscanner ); -void yypush_buffer_state (YY_BUFFER_STATE new_buffer ,yyscan_t yyscanner ); -void yypop_buffer_state (yyscan_t yyscanner ); +void yyrestart(FILE *input_file, yyscan_t yyscanner); +void yy_switch_to_buffer(YY_BUFFER_STATE new_buffer, yyscan_t yyscanner); +YY_BUFFER_STATE yy_create_buffer(FILE *file, int size, yyscan_t yyscanner); +void yy_delete_buffer(YY_BUFFER_STATE b, yyscan_t yyscanner); +void yy_flush_buffer(YY_BUFFER_STATE b, yyscan_t yyscanner); +void yypush_buffer_state(YY_BUFFER_STATE new_buffer, yyscan_t yyscanner); +void yypop_buffer_state(yyscan_t yyscanner); -static void yyensure_buffer_stack (yyscan_t yyscanner ); -static void yy_load_buffer_state (yyscan_t yyscanner ); -static void yy_init_buffer (YY_BUFFER_STATE b,FILE *file ,yyscan_t yyscanner ); +static void yyensure_buffer_stack(yyscan_t yyscanner); +static void yy_load_buffer_state(yyscan_t yyscanner); +static void yy_init_buffer(YY_BUFFER_STATE b, FILE *file, yyscan_t yyscanner); +#define YY_FLUSH_BUFFER yy_flush_buffer(YY_CURRENT_BUFFER, yyscanner) -#define YY_FLUSH_BUFFER yy_flush_buffer(YY_CURRENT_BUFFER ,yyscanner) +YY_BUFFER_STATE yy_scan_buffer(char *base, yy_size_t size, yyscan_t yyscanner); +YY_BUFFER_STATE yy_scan_string(const char *yy_str, yyscan_t yyscanner); +YY_BUFFER_STATE yy_scan_bytes(const char *bytes, yy_size_t len, yyscan_t yyscanner); -YY_BUFFER_STATE yy_scan_buffer (char *base,yy_size_t size ,yyscan_t yyscanner ); -YY_BUFFER_STATE yy_scan_string (yyconst char *yy_str ,yyscan_t yyscanner ); -YY_BUFFER_STATE yy_scan_bytes (yyconst char *bytes,int len ,yyscan_t yyscanner ); - -void *yyalloc (yy_size_t ,yyscan_t yyscanner ); -void *yyrealloc (void *,yy_size_t ,yyscan_t yyscanner ); -void yyfree (void * ,yyscan_t yyscanner ); +void *yyalloc(yy_size_t, yyscan_t yyscanner); +void *yyrealloc(void *, yy_size_t, yyscan_t yyscanner); +void yyfree(void *, yyscan_t yyscanner); #define yy_new_buffer yy_create_buffer - -#define yy_set_interactive(is_interactive) \ - { \ - if ( ! YY_CURRENT_BUFFER ){ \ - yyensure_buffer_stack (yyscanner); \ - YY_CURRENT_BUFFER_LVALUE = \ - yy_create_buffer(yyin,YY_BUF_SIZE ,yyscanner); \ - } \ - YY_CURRENT_BUFFER_LVALUE->yy_is_interactive = is_interactive; \ - } - -#define yy_set_bol(at_bol) \ - { \ - if ( ! YY_CURRENT_BUFFER ){\ - yyensure_buffer_stack (yyscanner); \ - YY_CURRENT_BUFFER_LVALUE = \ - yy_create_buffer(yyin,YY_BUF_SIZE ,yyscanner); \ - } \ - YY_CURRENT_BUFFER_LVALUE->yy_at_bol = at_bol; \ - } - +#define yy_set_interactive(is_interactive) \ + { \ + if (!YY_CURRENT_BUFFER) { \ + yyensure_buffer_stack(yyscanner); \ + YY_CURRENT_BUFFER_LVALUE = yy_create_buffer(yyin, YY_BUF_SIZE, yyscanner); \ + } \ + YY_CURRENT_BUFFER_LVALUE->yy_is_interactive = is_interactive; \ + } +#define yy_set_bol(at_bol) \ + { \ + if (!YY_CURRENT_BUFFER) { \ + yyensure_buffer_stack(yyscanner); \ + YY_CURRENT_BUFFER_LVALUE = yy_create_buffer(yyin, YY_BUF_SIZE, yyscanner); \ + } \ + YY_CURRENT_BUFFER_LVALUE->yy_at_bol = at_bol; \ + } #define YY_AT_BOL() (YY_CURRENT_BUFFER_LVALUE->yy_at_bol) /* Begin user sect3 */ -#define yywrap(n) 1 +#define yywrap(yyscanner) (/*CONSTCOND*/ 1) #define YY_SKIP_YYWRAP - -typedef unsigned char YY_CHAR; +typedef flex_uint8_t YY_CHAR; typedef int yy_state_type; #define yytext_ptr yytext_r -static yy_state_type yy_get_previous_state (yyscan_t yyscanner ); -static yy_state_type yy_try_NUL_trans (yy_state_type current_state ,yyscan_t yyscanner); -static int yy_get_next_buffer (yyscan_t yyscanner ); -static void yy_fatal_error (yyconst char msg[] ,yyscan_t yyscanner ); +static yy_state_type yy_get_previous_state(yyscan_t yyscanner); +static yy_state_type yy_try_NUL_trans(yy_state_type current_state, yyscan_t yyscanner); +static int yy_get_next_buffer(yyscan_t yyscanner); +static void yynoreturn yy_fatal_error(const char *msg, yyscan_t yyscanner); /* Done after the current pattern has been matched and before the * corresponding action - sets up yytext. */ -#define YY_DO_BEFORE_ACTION \ - yyg->yytext_ptr = yy_bp; \ - yyleng = (size_t) (yy_cp - yy_bp); \ - yyg->yy_hold_char = *yy_cp; \ - *yy_cp = '\0'; \ - yyg->yy_c_buf_p = yy_cp; - -#define YY_NUM_RULES 62 -#define YY_END_OF_BUFFER 63 +#define YY_DO_BEFORE_ACTION \ + yyg->yytext_ptr = yy_bp; \ + yyleng = (yy_size_t)(yy_cp - yy_bp); \ + yyg->yy_hold_char = *yy_cp; \ + *yy_cp = '\0'; \ + yyg->yy_c_buf_p = yy_cp; +#define YY_NUM_RULES 89 +#define YY_END_OF_BUFFER 90 /* This struct is not used in this scanner, but its presence is necessary. */ struct yy_trans_info - { - flex_int32_t yy_verify; - flex_int32_t yy_nxt; - }; -static yyconst flex_int16_t yy_accept[188] = - { 0, - 0, 0, 0, 0, 63, 61, 1, 2, 61, 61, - 61, 45, 46, 57, 55, 47, 56, 6, 58, 3, - 5, 52, 48, 54, 44, 44, 44, 44, 44, 44, - 44, 44, 44, 44, 44, 44, 44, 44, 44, 44, - 44, 44, 62, 51, 0, 59, 0, 60, 3, 0, - 49, 50, 53, 44, 44, 44, 44, 41, 44, 44, - 44, 44, 44, 44, 44, 44, 44, 44, 44, 44, - 44, 44, 44, 15, 44, 44, 44, 44, 44, 44, - 44, 44, 44, 44, 4, 22, 44, 44, 44, 44, - 44, 44, 44, 44, 44, 44, 44, 44, 44, 44, - - 44, 44, 44, 44, 44, 32, 44, 44, 44, 28, - 44, 44, 44, 44, 44, 44, 44, 44, 44, 19, - 33, 44, 44, 37, 44, 9, 11, 7, 44, 44, - 44, 20, 44, 8, 44, 44, 44, 24, 36, 44, - 44, 16, 44, 17, 44, 44, 44, 44, 44, 29, - 44, 44, 44, 44, 34, 44, 40, 14, 44, 44, - 44, 44, 44, 12, 44, 44, 44, 21, 30, 10, - 26, 44, 43, 38, 23, 44, 18, 44, 13, 27, - 25, 35, 39, 44, 42, 31, 0 - } ; - -static yyconst flex_int32_t yy_ec[256] = - { 0, - 1, 1, 1, 1, 1, 1, 1, 2, 2, 3, - 1, 2, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 2, 4, 5, 1, 1, 1, 1, 6, 7, - 8, 9, 10, 11, 12, 13, 14, 15, 15, 15, - 15, 15, 15, 15, 15, 15, 15, 1, 16, 17, - 18, 19, 1, 1, 20, 21, 22, 23, 24, 25, - 26, 27, 28, 29, 30, 31, 32, 33, 34, 35, - 29, 36, 37, 38, 39, 40, 41, 42, 43, 29, - 1, 1, 1, 1, 29, 1, 44, 45, 46, 47, - - 48, 49, 50, 51, 52, 29, 53, 54, 55, 56, - 57, 58, 29, 59, 60, 61, 62, 63, 64, 65, - 66, 29, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1 - } ; - -static yyconst flex_int32_t yy_meta[67] = - { 0, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 2, 1, 1, 1, 1, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2 - } ; - -static yyconst flex_int16_t yy_base[193] = - { 0, - 0, 0, 0, 0, 499, 500, 500, 500, 480, 492, - 490, 500, 500, 500, 500, 500, 480, 500, 500, 54, - 500, 52, 500, 476, 53, 57, 60, 59, 70, 91, - 61, 67, 75, 478, 87, 77, 98, 113, 109, 58, - 119, 111, 500, 500, 487, 500, 485, 500, 86, 475, - 500, 500, 500, 0, 474, 126, 121, 473, 115, 137, - 127, 151, 139, 153, 155, 150, 157, 161, 171, 172, - 165, 178, 189, 472, 180, 206, 183, 207, 203, 209, - 208, 195, 220, 228, 471, 462, 217, 231, 232, 233, - 236, 242, 246, 256, 248, 235, 243, 264, 257, 260, - - 266, 272, 275, 267, 285, 270, 278, 283, 296, 461, - 288, 295, 302, 301, 298, 306, 311, 307, 320, 460, - 459, 319, 321, 458, 323, 457, 456, 453, 326, 324, - 336, 452, 325, 449, 335, 342, 343, 448, 447, 348, - 352, 446, 366, 445, 363, 374, 373, 360, 375, 444, - 376, 377, 380, 388, 443, 391, 436, 434, 394, 392, - 400, 393, 398, 390, 417, 418, 407, 432, 411, 350, - 349, 423, 179, 177, 173, 424, 168, 421, 145, 143, - 99, 83, 74, 404, 73, 69, 500, 479, 481, 483, - 76, 75 - - } ; - -static yyconst flex_int16_t yy_def[193] = - { 0, - 187, 1, 188, 188, 187, 187, 187, 187, 187, 189, - 190, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 187, 187, 189, 187, 190, 187, 187, 187, - 187, 187, 187, 192, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 187, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 191, 191, 191, 191, - 191, 191, 191, 191, 191, 191, 0, 187, 187, 187, - 187, 187 - - } ; - -static yyconst flex_int16_t yy_nxt[567] = - { 0, - 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, - 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, - 26, 27, 28, 29, 30, 31, 32, 33, 34, 34, - 35, 34, 34, 36, 34, 37, 38, 39, 40, 41, - 42, 34, 34, 25, 26, 27, 28, 29, 30, 31, - 32, 33, 34, 35, 34, 34, 36, 34, 37, 38, - 39, 40, 41, 42, 34, 34, 50, 54, 49, 51, - 52, 54, 54, 54, 54, 54, 54, 55, 63, 59, - 57, 54, 64, 54, 54, 56, 60, 54, 54, 54, - 71, 54, 81, 61, 65, 62, 70, 54, 50, 58, - - 49, 54, 63, 59, 57, 54, 64, 72, 56, 74, - 60, 66, 54, 54, 71, 81, 61, 65, 62, 70, - 73, 67, 58, 54, 68, 54, 69, 54, 80, 54, - 72, 75, 74, 54, 66, 54, 76, 84, 82, 77, - 54, 54, 83, 73, 67, 88, 87, 68, 86, 69, - 78, 54, 80, 54, 75, 79, 89, 54, 90, 54, - 76, 84, 82, 77, 54, 54, 83, 54, 88, 54, - 87, 54, 86, 78, 91, 54, 92, 96, 79, 54, - 89, 90, 54, 93, 97, 54, 54, 54, 95, 94, - 98, 54, 54, 54, 54, 102, 99, 54, 91, 92, - - 103, 96, 104, 54, 100, 101, 93, 97, 107, 54, - 108, 95, 94, 98, 105, 106, 111, 54, 102, 99, - 54, 54, 54, 54, 103, 116, 104, 100, 101, 114, - 115, 54, 107, 108, 54, 113, 109, 105, 106, 111, - 112, 117, 54, 110, 119, 54, 54, 54, 116, 54, - 54, 118, 120, 114, 115, 123, 54, 54, 113, 109, - 54, 124, 54, 112, 122, 117, 110, 121, 119, 125, - 54, 54, 128, 129, 54, 118, 120, 126, 54, 123, - 54, 54, 127, 130, 54, 124, 54, 122, 131, 54, - 121, 132, 54, 125, 136, 128, 129, 54, 135, 54, - - 139, 126, 54, 138, 133, 127, 134, 130, 137, 54, - 54, 131, 54, 140, 132, 54, 54, 146, 136, 141, - 54, 54, 135, 144, 139, 54, 138, 133, 142, 134, - 143, 145, 137, 54, 54, 54, 140, 54, 54, 54, - 54, 146, 149, 141, 147, 154, 151, 144, 148, 54, - 54, 142, 150, 143, 145, 156, 54, 54, 152, 157, - 153, 155, 54, 54, 54, 149, 54, 147, 161, 154, - 151, 148, 159, 162, 54, 150, 158, 54, 160, 156, - 54, 152, 157, 153, 155, 163, 164, 54, 54, 54, - 54, 54, 161, 167, 54, 159, 166, 162, 168, 158, - - 170, 160, 54, 171, 54, 54, 54, 54, 54, 163, - 164, 165, 54, 169, 54, 172, 167, 174, 54, 176, - 166, 54, 168, 178, 170, 54, 179, 171, 173, 175, - 177, 54, 54, 186, 165, 54, 169, 54, 54, 172, - 180, 174, 182, 176, 185, 184, 54, 178, 54, 179, - 54, 173, 175, 177, 181, 183, 186, 54, 54, 54, - 54, 54, 54, 54, 180, 182, 54, 54, 185, 184, - 54, 54, 54, 54, 54, 54, 54, 181, 183, 43, - 43, 45, 45, 47, 47, 85, 54, 54, 54, 85, - 48, 46, 54, 53, 49, 48, 46, 44, 187, 5, - - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187 - } ; - -static yyconst flex_int16_t yy_chk[567] = - { 0, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 20, 25, 20, 22, - 22, 26, 40, 28, 27, 31, 192, 191, 28, 27, - 26, 32, 28, 186, 29, 25, 27, 185, 183, 33, - 32, 36, 40, 27, 28, 27, 31, 182, 49, 26, - - 49, 35, 28, 27, 26, 30, 28, 33, 25, 36, - 27, 29, 37, 181, 32, 40, 27, 28, 27, 31, - 35, 30, 26, 39, 30, 42, 30, 38, 39, 59, - 33, 37, 36, 41, 29, 57, 38, 42, 41, 38, - 56, 61, 41, 35, 30, 59, 57, 30, 56, 30, - 38, 60, 39, 63, 37, 38, 60, 180, 61, 179, - 38, 42, 41, 38, 66, 62, 41, 64, 59, 65, - 57, 67, 56, 38, 62, 68, 63, 66, 38, 71, - 60, 61, 177, 64, 66, 69, 70, 175, 65, 64, - 67, 174, 72, 173, 75, 71, 68, 77, 62, 63, - - 72, 66, 72, 73, 69, 70, 64, 66, 73, 82, - 75, 65, 64, 67, 72, 72, 77, 79, 71, 68, - 76, 78, 81, 80, 72, 82, 72, 69, 70, 80, - 81, 87, 73, 75, 83, 79, 76, 72, 72, 77, - 78, 83, 84, 76, 87, 88, 89, 90, 82, 96, - 91, 84, 88, 80, 81, 91, 92, 97, 79, 76, - 93, 92, 95, 78, 90, 83, 76, 89, 87, 93, - 94, 99, 96, 97, 100, 84, 88, 94, 98, 91, - 101, 104, 95, 98, 106, 92, 102, 90, 99, 103, - 89, 100, 107, 93, 104, 96, 97, 108, 103, 105, - - 107, 94, 111, 106, 101, 95, 102, 98, 105, 112, - 109, 99, 115, 108, 100, 114, 113, 115, 104, 109, - 116, 118, 103, 113, 107, 117, 106, 101, 111, 102, - 112, 114, 105, 122, 119, 123, 108, 125, 130, 133, - 129, 115, 118, 109, 116, 129, 122, 113, 117, 135, - 131, 111, 119, 112, 114, 131, 136, 137, 123, 133, - 125, 130, 140, 171, 170, 118, 141, 116, 140, 129, - 122, 117, 136, 141, 148, 119, 135, 145, 137, 131, - 143, 123, 133, 125, 130, 143, 145, 147, 146, 149, - 151, 152, 140, 148, 153, 136, 147, 141, 149, 135, - - 152, 137, 154, 153, 164, 156, 160, 162, 159, 143, - 145, 146, 163, 151, 161, 154, 148, 159, 184, 161, - 147, 167, 149, 163, 152, 169, 164, 153, 156, 160, - 162, 165, 166, 184, 146, 178, 151, 172, 176, 154, - 165, 159, 167, 161, 178, 176, 168, 163, 158, 164, - 157, 156, 160, 162, 166, 172, 184, 155, 150, 144, - 142, 139, 138, 134, 165, 167, 132, 128, 178, 176, - 127, 126, 124, 121, 120, 110, 86, 166, 172, 188, - 188, 189, 189, 190, 190, 85, 74, 58, 55, 50, - 47, 45, 34, 24, 17, 11, 10, 9, 5, 187, - - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187, 187, 187, 187, 187, - 187, 187, 187, 187, 187, 187 - } ; +{ + flex_int32_t yy_verify; + flex_int32_t yy_nxt; +}; +static const flex_int16_t yy_accept[271] = {0, + 0, + 0, + 0, + 0, + 90, + 88, + 1, + 2, + 88, + 88, + 88, + 66, + 67, + 84, + 82, + 70, + 83, + 6, + 85, + 3, + 5, + 75, + 71, + 77, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 68, + 69, + 89, + 74, + 0, + 86, + 0, + 87, + 0, + 3, + 72, + 73, + 76, + 81, + 81, + 81, + 49, + 81, + 48, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 51, + 79, + 81, + 81, + 81, + 81, + 81, + 15, + 23, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + + 81, + 81, + 81, + 81, + 81, + 81, + 4, + 22, + 50, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 33, + 81, + 81, + 81, + 81, + 81, + 78, + 81, + 81, + 81, + 81, + 81, + 29, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 81, + 19, + 34, + 81, + 81, + 36, + 81, + 9, + 81, + 11, + 81, + 7, + 81, + 81, + 81, + 81, + 20, + 81, + 81, + 8, + 81, + 81, + 81, + 81, + 25, + 81, + 56, + 80, + 81, + 81, + 40, + 81, + 81, + 81, + 81, + 16, + 81, + 17, + 81, + 37, + 64, + + 60, + 81, + 81, + 81, + 81, + 57, + 81, + 58, + 30, + 81, + 81, + 81, + 81, + 81, + 81, + 65, + 35, + 81, + 44, + 81, + 14, + 81, + 55, + 81, + 81, + 45, + 61, + 81, + 47, + 81, + 81, + 81, + 81, + 12, + 81, + 81, + 81, + 81, + 21, + 31, + 10, + 27, + 81, + 52, + 81, + 54, + 46, + 42, + 24, + 81, + 81, + 62, + 81, + 18, + 81, + 13, + 39, + 28, + 26, + 38, + 81, + 43, + 63, + 81, + 81, + 53, + 59, + 41, + 32, + 0}; + +static const YY_CHAR yy_ec[256] = {0, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 2, + 2, + 3, + 1, + 2, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 2, + 4, + 5, + 1, + 1, + 1, + 1, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 15, + 15, + 15, + 15, + 15, + 15, + 15, + 15, + 15, + 1, + 16, + 17, + 18, + 19, + 1, + 1, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 1, + 47, + 1, + 45, + 1, + 48, + 49, + 50, + 51, + + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 45, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1}; + +static const YY_CHAR yy_meta[73] = {0, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 2, + 1, + 1, + 1, + 1, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 1, + 1, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2}; + +static const flex_int16_t yy_base[276] = {0, + 0, + 0, + 0, + 0, + 741, + 742, + 742, + 742, + 722, + 734, + 732, + 742, + 742, + 742, + 742, + 742, + 742, + 742, + 742, + 60, + 742, + 58, + 742, + 719, + 59, + 63, + 64, + 65, + 66, + 107, + 67, + 116, + 73, + 71, + 721, + 68, + 113, + 128, + 108, + 84, + 135, + 134, + 161, + 177, + 176, + 742, + 742, + 742, + 742, + 730, + 742, + 728, + 742, + 718, + 95, + 742, + 742, + 742, + 0, + 717, + 133, + 145, + 162, + 716, + 183, + 137, + 148, + 185, + 195, + 197, + 198, + 196, + 211, + 212, + 204, + 203, + 226, + 235, + 206, + 230, + 283, + 715, + 229, + 237, + 269, + 233, + 255, + 714, + 200, + 242, + 264, + 272, + 258, + 262, + 276, + 234, + 270, + 265, + 275, + 287, + + 302, + 297, + 304, + 327, + 331, + 325, + 713, + 712, + 711, + 317, + 337, + 332, + 333, + 342, + 343, + 346, + 353, + 345, + 351, + 361, + 357, + 356, + 362, + 348, + 359, + 366, + 374, + 370, + 389, + 381, + 396, + 400, + 392, + 406, + 402, + 408, + 410, + 403, + 710, + 414, + 425, + 429, + 424, + 432, + 702, + 438, + 421, + 442, + 431, + 428, + 450, + 457, + 446, + 456, + 453, + 436, + 454, + 464, + 470, + 472, + 701, + 700, + 479, + 473, + 698, + 474, + 697, + 491, + 696, + 476, + 695, + 499, + 502, + 483, + 503, + 694, + 485, + 501, + 692, + 484, + 513, + 515, + 516, + 690, + 527, + 689, + 688, + 517, + 521, + 545, + 530, + 522, + 542, + 546, + 687, + 549, + 685, + 551, + 683, + 682, + + 680, + 558, + 555, + 561, + 556, + 679, + 564, + 678, + 677, + 562, + 568, + 584, + 567, + 574, + 587, + 672, + 669, + 591, + 668, + 595, + 667, + 590, + 655, + 592, + 589, + 652, + 651, + 596, + 647, + 608, + 604, + 612, + 618, + 610, + 617, + 623, + 619, + 624, + 620, + 563, + 534, + 523, + 638, + 386, + 635, + 363, + 315, + 303, + 279, + 639, + 640, + 266, + 641, + 263, + 649, + 244, + 238, + 205, + 202, + 168, + 662, + 155, + 136, + 666, + 650, + 101, + 88, + 85, + 80, + 742, + 718, + 720, + 722, + 92, + 88}; + +static const flex_int16_t yy_def[276] = {0, + 270, + 1, + 271, + 271, + 270, + 270, + 270, + 270, + 270, + 272, + 273, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 270, + 270, + 270, + 270, + 272, + 270, + 273, + 270, + 270, + 270, + 270, + 270, + 270, + 275, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + + 274, + 274, + 274, + 274, + 274, + 274, + 270, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 274, + 0, + 270, + 270, + 270, + 270, + 270}; + +static const flex_int16_t yy_nxt[815] = {0, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 35, + 37, + 38, + 39, + 35, + 40, + 41, + 42, + 43, + 44, + 45, + 35, + 35, + 35, + 46, + 47, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 35, + 37, + 38, + 39, + 35, + 40, + 41, + 42, + 43, + 44, + 45, + 35, + 35, + 54, + 59, + 55, + 56, + 57, + 59, + 59, + 59, + 59, + 59, + 59, + 65, + 69, + 59, + 63, + 59, + 70, + 59, + 66, + 61, + 71, + 60, + 59, + 85, + 62, + 67, + 59, + 59, + + 68, + 72, + 59, + 78, + 84, + 81, + 64, + 54, + 73, + 55, + 82, + 65, + 69, + 83, + 63, + 59, + 70, + 91, + 66, + 61, + 71, + 59, + 59, + 85, + 62, + 67, + 74, + 59, + 68, + 72, + 59, + 78, + 84, + 81, + 64, + 79, + 73, + 75, + 82, + 80, + 76, + 83, + 59, + 77, + 90, + 91, + 86, + 59, + 59, + 59, + 59, + 59, + 87, + 96, + 74, + 108, + 112, + 97, + 92, + 59, + 88, + 93, + 59, + 79, + 89, + 75, + 109, + 80, + 76, + 59, + 98, + 77, + 90, + 94, + 86, + 59, + 59, + 99, + 95, + 113, + 87, + 96, + 59, + 108, + 112, + 97, + 92, + 110, + 88, + 93, + 59, + 59, + 89, + 100, + 109, + 101, + 102, + 59, + 98, + 59, + + 103, + 94, + 105, + 106, + 104, + 99, + 95, + 113, + 114, + 59, + 59, + 59, + 59, + 111, + 59, + 110, + 59, + 59, + 59, + 59, + 59, + 100, + 141, + 101, + 102, + 59, + 59, + 116, + 103, + 119, + 105, + 106, + 104, + 115, + 117, + 118, + 114, + 123, + 120, + 124, + 59, + 111, + 122, + 59, + 59, + 121, + 127, + 59, + 59, + 59, + 141, + 59, + 59, + 134, + 149, + 116, + 59, + 119, + 59, + 125, + 128, + 115, + 117, + 118, + 135, + 123, + 120, + 124, + 126, + 59, + 122, + 139, + 59, + 121, + 127, + 142, + 59, + 59, + 59, + 59, + 59, + 134, + 149, + 59, + 59, + 140, + 59, + 125, + 128, + 59, + 59, + 146, + 135, + 59, + 143, + 147, + 126, + 59, + 136, + 139, + + 137, + 59, + 144, + 142, + 151, + 129, + 138, + 130, + 148, + 152, + 145, + 59, + 150, + 140, + 153, + 131, + 59, + 59, + 59, + 146, + 132, + 133, + 143, + 147, + 154, + 156, + 136, + 155, + 137, + 59, + 144, + 59, + 151, + 129, + 138, + 130, + 148, + 152, + 145, + 59, + 150, + 59, + 153, + 131, + 160, + 59, + 59, + 59, + 132, + 133, + 157, + 59, + 154, + 156, + 158, + 155, + 59, + 59, + 161, + 59, + 59, + 164, + 59, + 159, + 163, + 59, + 165, + 59, + 162, + 166, + 59, + 59, + 160, + 59, + 167, + 59, + 59, + 59, + 157, + 175, + 59, + 174, + 158, + 168, + 59, + 169, + 161, + 172, + 59, + 164, + 176, + 159, + 163, + 173, + 165, + 59, + 162, + 166, + 170, + 171, + + 59, + 178, + 167, + 59, + 179, + 177, + 59, + 175, + 181, + 174, + 59, + 168, + 180, + 169, + 59, + 172, + 59, + 59, + 176, + 182, + 59, + 173, + 59, + 183, + 59, + 184, + 170, + 171, + 59, + 178, + 185, + 187, + 179, + 177, + 186, + 59, + 181, + 188, + 59, + 59, + 180, + 189, + 59, + 59, + 190, + 59, + 59, + 182, + 191, + 192, + 59, + 183, + 59, + 184, + 193, + 194, + 59, + 196, + 185, + 187, + 59, + 198, + 186, + 197, + 59, + 188, + 199, + 59, + 59, + 189, + 59, + 59, + 190, + 200, + 205, + 203, + 191, + 192, + 59, + 195, + 201, + 202, + 193, + 194, + 59, + 196, + 59, + 59, + 59, + 198, + 59, + 197, + 204, + 59, + 199, + 206, + 208, + 59, + 59, + 59, + + 207, + 200, + 205, + 203, + 209, + 59, + 210, + 195, + 201, + 202, + 213, + 211, + 212, + 59, + 214, + 59, + 59, + 59, + 215, + 219, + 204, + 217, + 218, + 206, + 208, + 216, + 221, + 59, + 207, + 59, + 59, + 59, + 209, + 220, + 210, + 59, + 59, + 59, + 213, + 211, + 212, + 59, + 214, + 222, + 59, + 230, + 215, + 219, + 59, + 217, + 218, + 223, + 224, + 216, + 221, + 226, + 59, + 225, + 227, + 59, + 59, + 220, + 231, + 59, + 228, + 59, + 229, + 232, + 233, + 59, + 59, + 222, + 59, + 230, + 234, + 59, + 59, + 59, + 59, + 223, + 224, + 59, + 59, + 226, + 237, + 225, + 227, + 239, + 59, + 238, + 231, + 241, + 228, + 236, + 229, + 232, + 233, + 235, + 59, + 243, + + 240, + 59, + 234, + 59, + 59, + 59, + 59, + 242, + 250, + 59, + 59, + 244, + 237, + 248, + 245, + 239, + 251, + 238, + 59, + 241, + 247, + 236, + 59, + 253, + 59, + 235, + 59, + 243, + 240, + 246, + 249, + 59, + 59, + 59, + 59, + 242, + 250, + 59, + 59, + 244, + 257, + 248, + 245, + 255, + 251, + 252, + 258, + 256, + 247, + 59, + 254, + 253, + 59, + 59, + 59, + 59, + 259, + 246, + 249, + 261, + 260, + 59, + 265, + 59, + 59, + 59, + 59, + 262, + 257, + 59, + 264, + 255, + 266, + 252, + 258, + 256, + 59, + 263, + 254, + 269, + 59, + 59, + 59, + 59, + 259, + 267, + 59, + 261, + 260, + 268, + 265, + 59, + 59, + 59, + 59, + 262, + 59, + 59, + 264, + 59, + + 266, + 59, + 59, + 59, + 59, + 263, + 59, + 269, + 59, + 59, + 59, + 59, + 59, + 267, + 59, + 59, + 59, + 268, + 48, + 48, + 50, + 50, + 52, + 52, + 59, + 59, + 59, + 107, + 59, + 59, + 59, + 59, + 107, + 53, + 51, + 59, + 58, + 53, + 51, + 49, + 270, + 5, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270}; + +static const flex_int16_t yy_chk[815] = {0, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 20, + 25, + 20, + 22, + 22, + 26, + 27, + 28, + 29, + 31, + 36, + 27, + 28, + 34, + 26, + 33, + 28, + 275, + 27, + 25, + 28, + 274, + 269, + 36, + 25, + 27, + 40, + 268, + + 27, + 28, + 267, + 31, + 34, + 33, + 26, + 55, + 29, + 55, + 33, + 27, + 28, + 33, + 26, + 266, + 28, + 40, + 27, + 25, + 28, + 30, + 39, + 36, + 25, + 27, + 30, + 37, + 27, + 28, + 32, + 31, + 34, + 33, + 26, + 32, + 29, + 30, + 33, + 32, + 30, + 33, + 38, + 30, + 39, + 40, + 37, + 61, + 42, + 41, + 263, + 66, + 37, + 42, + 30, + 61, + 66, + 42, + 41, + 62, + 38, + 41, + 67, + 32, + 38, + 30, + 62, + 32, + 30, + 262, + 42, + 30, + 39, + 41, + 37, + 43, + 63, + 42, + 41, + 67, + 37, + 42, + 260, + 61, + 66, + 42, + 41, + 63, + 38, + 41, + 45, + 44, + 38, + 43, + 62, + 43, + 44, + 65, + 42, + 68, + + 44, + 41, + 45, + 45, + 44, + 42, + 41, + 67, + 68, + 69, + 72, + 70, + 71, + 65, + 89, + 63, + 259, + 76, + 75, + 258, + 79, + 43, + 89, + 43, + 44, + 73, + 74, + 70, + 44, + 72, + 45, + 45, + 44, + 69, + 70, + 71, + 68, + 75, + 73, + 76, + 77, + 65, + 74, + 83, + 80, + 73, + 79, + 86, + 96, + 78, + 89, + 84, + 257, + 83, + 96, + 70, + 90, + 72, + 256, + 77, + 80, + 69, + 70, + 71, + 84, + 75, + 73, + 76, + 78, + 87, + 74, + 86, + 93, + 73, + 79, + 90, + 94, + 254, + 91, + 98, + 252, + 83, + 96, + 85, + 97, + 87, + 92, + 77, + 80, + 99, + 95, + 93, + 84, + 249, + 91, + 94, + 78, + 81, + 85, + 86, + + 85, + 100, + 92, + 90, + 98, + 81, + 85, + 81, + 95, + 99, + 92, + 102, + 97, + 87, + 100, + 81, + 101, + 248, + 103, + 93, + 81, + 81, + 91, + 94, + 101, + 103, + 85, + 102, + 85, + 247, + 92, + 110, + 98, + 81, + 85, + 81, + 95, + 99, + 92, + 106, + 97, + 104, + 100, + 81, + 110, + 105, + 112, + 113, + 81, + 81, + 104, + 111, + 101, + 103, + 105, + 102, + 114, + 115, + 111, + 118, + 116, + 114, + 124, + 106, + 113, + 119, + 115, + 117, + 112, + 116, + 122, + 121, + 110, + 125, + 117, + 120, + 123, + 246, + 104, + 124, + 126, + 123, + 105, + 118, + 128, + 119, + 111, + 121, + 127, + 114, + 125, + 106, + 113, + 122, + 115, + 130, + 112, + 116, + 120, + 120, + + 244, + 127, + 117, + 129, + 128, + 126, + 133, + 124, + 130, + 123, + 131, + 118, + 129, + 119, + 132, + 121, + 135, + 138, + 125, + 131, + 134, + 122, + 136, + 132, + 137, + 133, + 120, + 120, + 140, + 127, + 134, + 136, + 128, + 126, + 135, + 147, + 130, + 137, + 143, + 141, + 129, + 138, + 150, + 142, + 140, + 149, + 144, + 131, + 141, + 142, + 156, + 132, + 146, + 133, + 143, + 144, + 148, + 147, + 134, + 136, + 153, + 149, + 135, + 148, + 151, + 137, + 150, + 155, + 157, + 138, + 154, + 152, + 140, + 151, + 156, + 154, + 141, + 142, + 158, + 146, + 152, + 153, + 143, + 144, + 159, + 147, + 160, + 164, + 166, + 149, + 170, + 148, + 155, + 163, + 150, + 157, + 159, + 174, + 180, + 177, + + 158, + 151, + 156, + 154, + 160, + 168, + 163, + 146, + 152, + 153, + 168, + 164, + 166, + 172, + 170, + 178, + 173, + 175, + 172, + 177, + 155, + 174, + 175, + 157, + 159, + 173, + 180, + 181, + 158, + 182, + 183, + 188, + 160, + 178, + 163, + 189, + 192, + 242, + 168, + 164, + 166, + 185, + 170, + 181, + 191, + 192, + 172, + 177, + 241, + 174, + 175, + 182, + 183, + 173, + 180, + 188, + 193, + 185, + 189, + 190, + 194, + 178, + 193, + 196, + 190, + 198, + 191, + 194, + 196, + 203, + 205, + 181, + 202, + 192, + 198, + 204, + 210, + 240, + 207, + 182, + 183, + 213, + 211, + 188, + 204, + 185, + 189, + 207, + 214, + 205, + 193, + 211, + 190, + 203, + 191, + 194, + 196, + 202, + 212, + 213, + + 210, + 215, + 198, + 225, + 222, + 218, + 224, + 212, + 225, + 220, + 228, + 214, + 204, + 222, + 215, + 207, + 228, + 205, + 231, + 211, + 220, + 203, + 230, + 231, + 234, + 202, + 232, + 213, + 210, + 218, + 224, + 235, + 233, + 237, + 239, + 212, + 225, + 236, + 238, + 214, + 235, + 222, + 215, + 233, + 228, + 230, + 236, + 234, + 220, + 245, + 232, + 231, + 243, + 250, + 251, + 253, + 237, + 218, + 224, + 243, + 238, + 229, + 253, + 255, + 265, + 227, + 226, + 245, + 235, + 223, + 251, + 233, + 255, + 230, + 236, + 234, + 261, + 250, + 232, + 265, + 264, + 221, + 219, + 217, + 237, + 261, + 216, + 243, + 238, + 264, + 253, + 209, + 208, + 206, + 201, + 245, + 200, + 199, + 251, + 197, + + 255, + 195, + 187, + 186, + 184, + 250, + 179, + 265, + 176, + 171, + 169, + 167, + 165, + 261, + 162, + 161, + 145, + 264, + 271, + 271, + 272, + 272, + 273, + 273, + 139, + 109, + 108, + 107, + 88, + 82, + 64, + 60, + 54, + 52, + 50, + 35, + 24, + 11, + 10, + 9, + 5, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270, + 270}; /* The intent behind this definition is that it'll catch * any uses of REJECT which flex missed. @@ -645,8 +3209,8 @@ static yyconst flex_int16_t yy_chk[567] = #line 1 "lex_sql.l" #line 28 "lex_sql.l" -#include -#include +#include +#include /** * flex 代码包含三个部分，使用 %% 分隔 @@ -660,13 +3224,16 @@ static yyconst flex_int16_t yy_chk[567] = #include "yacc_sql.hpp" #ifndef register -#define register -#endif // register +#define register +#endif // register -extern int atoi(); +extern int atoi(); extern double atof(); -#define RETURN_TOKEN(token) LOG_DEBUG("%s", #token);return token +#define RETURN_TOKEN(token) \ + LOG_DEBUG("%s", #token); \ + return token +#line 763 "lex_sql.cpp" /* Prevent the need for linking with -lfl */ #define YY_NO_INPUT 1 /* 不区分大小写 */ @@ -675,7 +3242,7 @@ extern double atof(); /* 1. 匹配的规则长的优先 */ /* 2. 写在最前面的优先 */ /* yylval 就可以认为是 yacc 中 %union 定义的结构体(union 结构) */ -#line 679 "lex_sql.cpp" +#line 772 "lex_sql.cpp" #define INITIAL 0 #define STR 1 @@ -694,124 +3261,136 @@ extern double atof(); /* Holds the entire state of the reentrant scanner. */ struct yyguts_t - { +{ + + /* User-defined. Not touched by flex. */ + YY_EXTRA_TYPE yyextra_r; + + /* The rest are the same as the globals declared in the non-reentrant scanner. */ + FILE *yyin_r, *yyout_r; + size_t yy_buffer_stack_top; /**< index of top of stack. */ + size_t yy_buffer_stack_max; /**< capacity of stack. */ + YY_BUFFER_STATE *yy_buffer_stack; /**< Stack as an array. */ + char yy_hold_char; + yy_size_t yy_n_chars; + yy_size_t yyleng_r; + char *yy_c_buf_p; + int yy_init; + int yy_start; + int yy_did_buffer_switch_on_eof; + int yy_start_stack_ptr; + int yy_start_stack_depth; + int *yy_start_stack; + yy_state_type yy_last_accepting_state; + char *yy_last_accepting_cpos; - /* User-defined. Not touched by flex. */ - YY_EXTRA_TYPE yyextra_r; + int yylineno_r; + int yy_flex_debug_r; - /* The rest are the same as the globals declared in the non-reentrant scanner. */ - FILE *yyin_r, *yyout_r; - size_t yy_buffer_stack_top; /**< index of top of stack. */ - size_t yy_buffer_stack_max; /**< capacity of stack. */ - YY_BUFFER_STATE * yy_buffer_stack; /**< Stack as an array. */ - char yy_hold_char; - int yy_n_chars; - int yyleng_r; - char *yy_c_buf_p; - int yy_init; - int yy_start; - int yy_did_buffer_switch_on_eof; - int yy_start_stack_ptr; - int yy_start_stack_depth; - int *yy_start_stack; - yy_state_type yy_last_accepting_state; - char* yy_last_accepting_cpos; + char *yytext_r; + int yy_more_flag; + int yy_more_len; - int yylineno_r; - int yy_flex_debug_r; + YYSTYPE *yylval_r; - char *yytext_r; - int yy_more_flag; - int yy_more_len; + YYLTYPE *yylloc_r; - YYSTYPE * yylval_r; +}; /* end struct yyguts_t */ - YYLTYPE * yylloc_r; +static int yy_init_globals(yyscan_t yyscanner); - }; /* end struct yyguts_t */ +/* This must go here because YYSTYPE and YYLTYPE are included + * from bison output in section 1.*/ +#define yylval yyg->yylval_r -static int yy_init_globals (yyscan_t yyscanner ); +#define yylloc yyg->yylloc_r - /* This must go here because YYSTYPE and YYLTYPE are included - * from bison output in section 1.*/ - # define yylval yyg->yylval_r - - # define yylloc yyg->yylloc_r - -int yylex_init (yyscan_t* scanner); +int yylex_init(yyscan_t *scanner); -int yylex_init_extra (YY_EXTRA_TYPE user_defined,yyscan_t* scanner); +int yylex_init_extra(YY_EXTRA_TYPE user_defined, yyscan_t *scanner); /* Accessor methods to globals. These are made visible to non-reentrant scanners for convenience. */ -int yylex_destroy (yyscan_t yyscanner ); +int yylex_destroy(yyscan_t yyscanner); + +int yyget_debug(yyscan_t yyscanner); + +void yyset_debug(int debug_flag, yyscan_t yyscanner); + +YY_EXTRA_TYPE yyget_extra(yyscan_t yyscanner); -int yyget_debug (yyscan_t yyscanner ); +void yyset_extra(YY_EXTRA_TYPE user_defined, yyscan_t yyscanner); -void yyset_debug (int debug_flag ,yyscan_t yyscanner ); +FILE *yyget_in(yyscan_t yyscanner); -YY_EXTRA_TYPE yyget_extra (yyscan_t yyscanner ); +void yyset_in(FILE *_in_str, yyscan_t yyscanner); -void yyset_extra (YY_EXTRA_TYPE user_defined ,yyscan_t yyscanner ); +FILE *yyget_out(yyscan_t yyscanner); -FILE *yyget_in (yyscan_t yyscanner ); +void yyset_out(FILE *_out_str, yyscan_t yyscanner); -void yyset_in (FILE * in_str ,yyscan_t yyscanner ); +yy_size_t yyget_leng(yyscan_t yyscanner); -FILE *yyget_out (yyscan_t yyscanner ); +char *yyget_text(yyscan_t yyscanner); -void yyset_out (FILE * out_str ,yyscan_t yyscanner ); +int yyget_lineno(yyscan_t yyscanner); -int yyget_leng (yyscan_t yyscanner ); +void yyset_lineno(int _line_number, yyscan_t yyscanner); -char *yyget_text (yyscan_t yyscanner ); +int yyget_column(yyscan_t yyscanner); -int yyget_lineno (yyscan_t yyscanner ); +void yyset_column(int _column_no, yyscan_t yyscanner); -void yyset_lineno (int line_number ,yyscan_t yyscanner ); +YYSTYPE *yyget_lval(yyscan_t yyscanner); -YYSTYPE * yyget_lval (yyscan_t yyscanner ); +void yyset_lval(YYSTYPE *yylval_param, yyscan_t yyscanner); -void yyset_lval (YYSTYPE * yylval_param ,yyscan_t yyscanner ); +YYLTYPE *yyget_lloc(yyscan_t yyscanner); + +void yyset_lloc(YYLTYPE *yylloc_param, yyscan_t yyscanner); - YYLTYPE *yyget_lloc (yyscan_t yyscanner ); - - void yyset_lloc (YYLTYPE * yylloc_param ,yyscan_t yyscanner ); - /* Macros after this point can all be overridden by user definitions in * section 1. */ #ifndef YY_SKIP_YYWRAP #ifdef __cplusplus -extern "C" int yywrap (yyscan_t yyscanner ); +extern "C" int yywrap(yyscan_t yyscanner); #else -extern int yywrap (yyscan_t yyscanner ); +extern int yywrap(yyscan_t yyscanner); +#endif #endif + +#ifndef YY_NO_UNPUT + #endif #ifndef yytext_ptr -static void yy_flex_strncpy (char *,yyconst char *,int ,yyscan_t yyscanner); +static void yy_flex_strncpy(char *, const char *, int, yyscan_t yyscanner); #endif #ifdef YY_NEED_STRLEN -static int yy_flex_strlen (yyconst char * ,yyscan_t yyscanner); +static int yy_flex_strlen(const char *, yyscan_t yyscanner); #endif #ifndef YY_NO_INPUT - #ifdef __cplusplus -static int yyinput (yyscan_t yyscanner ); +static int yyinput(yyscan_t yyscanner); #else -static int input (yyscan_t yyscanner ); +static int input(yyscan_t yyscanner); #endif #endif /* Amount of stuff to slurp up with each read. */ #ifndef YY_READ_BUF_SIZE +#ifdef __ia64__ +/* On IA-64, the buffer size is 16k, not 8k */ +#define YY_READ_BUF_SIZE 16384 +#else #define YY_READ_BUF_SIZE 8192 +#endif /* __ia64__ */ #endif /* Copy whatever the last rule matched to the standard output. */ @@ -819,42 +3398,38 @@ static int input (yyscan_t yyscanner ); /* This used to be an fputs(), but since the string might contain NUL's, * we now use fwrite(). */ -#define ECHO fwrite( yytext, yyleng, 1, yyout ) +#define ECHO \ + do { \ + if (fwrite(yytext, (size_t)yyleng, 1, yyout)) {} \ + } while (0) #endif /* Gets input and stuffs it into "buf". number of characters read, or YY_NULL, * is returned in "result". */ #ifndef YY_INPUT -#define YY_INPUT(buf,result,max_size) \ - if ( YY_CURRENT_BUFFER_LVALUE->yy_is_interactive ) \ - { \ - int c = '*'; \ - int n; \ - for ( n = 0; n < max_size && \ - (c = getc( yyin )) != EOF && c != '\n'; ++n ) \ - buf[n] = (char) c; \ - if ( c == '\n' ) \ - buf[n++] = (char) c; \ - if ( c == EOF && ferror( yyin ) ) \ - YY_FATAL_ERROR( "input in flex scanner failed" ); \ - result = n; \ - } \ - else \ - { \ - errno=0; \ - while ( (result = fread(buf, 1, max_size, yyin))==0 && ferror(yyin)) \ - { \ - if( errno != EINTR) \ - { \ - YY_FATAL_ERROR( "input in flex scanner failed" ); \ - break; \ - } \ - errno=0; \ - clearerr(yyin); \ - } \ - }\ -\ +#define YY_INPUT(buf, result, max_size) \ + if (YY_CURRENT_BUFFER_LVALUE->yy_is_interactive) { \ + int c = '*'; \ + yy_size_t n; \ + for (n = 0; n < max_size && (c = getc(yyin)) != EOF && c != '\n'; ++n) \ + buf[n] = (char)c; \ + if (c == '\n') \ + buf[n++] = (char)c; \ + if (c == EOF && ferror(yyin)) \ + YY_FATAL_ERROR("input in flex scanner failed"); \ + result = n; \ + } else { \ + errno = 0; \ + while ((result = (int)fread(buf, 1, (yy_size_t)max_size, yyin)) == 0 && ferror(yyin)) { \ + if (errno != EINTR) { \ + YY_FATAL_ERROR("input in flex scanner failed"); \ + break; \ + } \ + errno = 0; \ + clearerr(yyin); \ + } \ + } #endif @@ -873,7 +3448,7 @@ static int input (yyscan_t yyscanner ); /* Report a fatal error. */ #ifndef YY_FATAL_ERROR -#define YY_FATAL_ERROR(msg) yy_fatal_error( msg , yyscanner) +#define YY_FATAL_ERROR(msg) yy_fatal_error(msg, yyscanner) #endif /* end tables serialization structures and prototypes */ @@ -884,11 +3459,9 @@ static int input (yyscan_t yyscanner ); #ifndef YY_DECL #define YY_DECL_IS_OURS 1 -extern int yylex \ - (YYSTYPE * yylval_param,YYLTYPE * yylloc_param ,yyscan_t yyscanner); +extern int yylex(YYSTYPE *yylval_param, YYLTYPE *yylloc_param, yyscan_t yyscanner); -#define YY_DECL int yylex \ - (YYSTYPE * yylval_param, YYLTYPE * yylloc_param , yyscan_t yyscanner) +#define YY_DECL int yylex(YYSTYPE *yylval_param, YYLTYPE *yylloc_param, yyscan_t yyscanner) #endif /* !YY_DECL */ /* Code executed at the beginning of each rule, after yytext and yyleng @@ -900,547 +3473,586 @@ extern int yylex \ /* Code executed at the end of each rule. */ #ifndef YY_BREAK -#define YY_BREAK break; +#define YY_BREAK /*LINTED*/ break; #endif -#define YY_RULE_SETUP \ - YY_USER_ACTION +#define YY_RULE_SETUP YY_USER_ACTION /** The main scanner function which does all the work. */ YY_DECL { - register yy_state_type yy_current_state; - register char *yy_cp, *yy_bp; - register int yy_act; - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; + yy_state_type yy_current_state; + char *yy_cp, *yy_bp; + int yy_act; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; -#line 75 "lex_sql.l" + yylval = yylval_param; + yylloc = yylloc_param; -#line 922 "lex_sql.cpp" + if (!yyg->yy_init) { + yyg->yy_init = 1; - yylval = yylval_param; +#ifdef YY_USER_INIT + YY_USER_INIT; +#endif - yylloc = yylloc_param; + if (!yyg->yy_start) + yyg->yy_start = 1; /* first start state */ - if ( !yyg->yy_init ) - { - yyg->yy_init = 1; + if (!yyin) + yyin = stdin; -#ifdef YY_USER_INIT - YY_USER_INIT; -#endif + if (!yyout) + yyout = stdout; - if ( ! yyg->yy_start ) - yyg->yy_start = 1; /* first start state */ - - if ( ! yyin ) - yyin = stdin; - - if ( ! yyout ) - yyout = stdout; - - if ( ! YY_CURRENT_BUFFER ) { - yyensure_buffer_stack (yyscanner); - YY_CURRENT_BUFFER_LVALUE = - yy_create_buffer(yyin,YY_BUF_SIZE ,yyscanner); - } - - yy_load_buffer_state(yyscanner ); - } - - while ( 1 ) /* loops until end-of-file is reached */ - { - yy_cp = yyg->yy_c_buf_p; - - /* Support of yytext. */ - *yy_cp = yyg->yy_hold_char; - - /* yy_bp points to the position in yy_ch_buf of the start of - * the current run. - */ - yy_bp = yy_cp; - - yy_current_state = yyg->yy_start; -yy_match: - do - { - register YY_CHAR yy_c = yy_ec[YY_SC_TO_UI(*yy_cp)]; - if ( yy_accept[yy_current_state] ) - { - yyg->yy_last_accepting_state = yy_current_state; - yyg->yy_last_accepting_cpos = yy_cp; - } - while ( yy_chk[yy_base[yy_current_state] + yy_c] != yy_current_state ) - { - yy_current_state = (int) yy_def[yy_current_state]; - if ( yy_current_state >= 188 ) - yy_c = yy_meta[(unsigned int) yy_c]; - } - yy_current_state = yy_nxt[yy_base[yy_current_state] + (unsigned int) yy_c]; - ++yy_cp; - } - while ( yy_base[yy_current_state] != 500 ); - -yy_find_action: - yy_act = yy_accept[yy_current_state]; - if ( yy_act == 0 ) - { /* have to back up */ - yy_cp = yyg->yy_last_accepting_cpos; - yy_current_state = yyg->yy_last_accepting_state; - yy_act = yy_accept[yy_current_state]; - } - - YY_DO_BEFORE_ACTION; - -do_action: /* This label is used only to access EOF actions. */ - - switch ( yy_act ) - { /* beginning of action switch */ - case 0: /* must back up */ - /* undo the effects of YY_DO_BEFORE_ACTION */ - *yy_cp = yyg->yy_hold_char; - yy_cp = yyg->yy_last_accepting_cpos; - yy_current_state = yyg->yy_last_accepting_state; - goto yy_find_action; - -case 1: -YY_RULE_SETUP + if (!YY_CURRENT_BUFFER) { + yyensure_buffer_stack(yyscanner); + YY_CURRENT_BUFFER_LVALUE = yy_create_buffer(yyin, YY_BUF_SIZE, yyscanner); + } + + yy_load_buffer_state(yyscanner); + } + + { +#line 75 "lex_sql.l" + +#line 1058 "lex_sql.cpp" + + while (/*CONSTCOND*/ 1) /* loops until end-of-file is reached */ + { + yy_cp = yyg->yy_c_buf_p; + + /* Support of yytext. */ + *yy_cp = yyg->yy_hold_char; + + /* yy_bp points to the position in yy_ch_buf of the start of + * the current run. + */ + yy_bp = yy_cp; + + yy_current_state = yyg->yy_start; + yy_match: + do { + YY_CHAR yy_c = yy_ec[YY_SC_TO_UI(*yy_cp)]; + if (yy_accept[yy_current_state]) { + yyg->yy_last_accepting_state = yy_current_state; + yyg->yy_last_accepting_cpos = yy_cp; + } + while (yy_chk[yy_base[yy_current_state] + yy_c] != yy_current_state) { + yy_current_state = (int)yy_def[yy_current_state]; + if (yy_current_state >= 271) + yy_c = yy_meta[yy_c]; + } + yy_current_state = yy_nxt[yy_base[yy_current_state] + yy_c]; + ++yy_cp; + } while (yy_base[yy_current_state] != 742); + + yy_find_action: + yy_act = yy_accept[yy_current_state]; + if (yy_act == 0) { /* have to back up */ + yy_cp = yyg->yy_last_accepting_cpos; + yy_current_state = yyg->yy_last_accepting_state; + yy_act = yy_accept[yy_current_state]; + } + + YY_DO_BEFORE_ACTION; + + do_action: /* This label is used only to access EOF actions. */ + + switch (yy_act) { /* beginning of action switch */ + case 0: /* must back up */ + /* undo the effects of YY_DO_BEFORE_ACTION */ + *yy_cp = yyg->yy_hold_char; + yy_cp = yyg->yy_last_accepting_cpos; + yy_current_state = yyg->yy_last_accepting_state; + goto yy_find_action; + + case 1: YY_RULE_SETUP #line 77 "lex_sql.l" -// ignore whitespace - YY_BREAK -case 2: -/* rule 2 can match eol */ -YY_RULE_SETUP + // ignore whitespace + YY_BREAK + case 2: + /* rule 2 can match eol */ + YY_RULE_SETUP #line 78 "lex_sql.l" -; - YY_BREAK -case 3: -YY_RULE_SETUP + ; + YY_BREAK + case 3: YY_RULE_SETUP #line 80 "lex_sql.l" -yylval->number=atoi(yytext); RETURN_TOKEN(NUMBER); - YY_BREAK -case 4: -YY_RULE_SETUP + yylval->number = atoi(yytext); + RETURN_TOKEN(NUMBER); + YY_BREAK + case 4: YY_RULE_SETUP #line 81 "lex_sql.l" -yylval->floats=(float)(atof(yytext)); RETURN_TOKEN(FLOAT); - YY_BREAK -case 5: -YY_RULE_SETUP + yylval->floats = (float)(atof(yytext)); + RETURN_TOKEN(FLOAT); + YY_BREAK + case 5: YY_RULE_SETUP #line 83 "lex_sql.l" -RETURN_TOKEN(SEMICOLON); - YY_BREAK -case 6: -YY_RULE_SETUP + RETURN_TOKEN(SEMICOLON); + YY_BREAK + case 6: YY_RULE_SETUP #line 84 "lex_sql.l" -RETURN_TOKEN(DOT); - YY_BREAK -case 7: -YY_RULE_SETUP + RETURN_TOKEN(DOT); + YY_BREAK + case 7: YY_RULE_SETUP #line 85 "lex_sql.l" -RETURN_TOKEN(EXIT); - YY_BREAK -case 8: -YY_RULE_SETUP + RETURN_TOKEN(EXIT); + YY_BREAK + case 8: YY_RULE_SETUP #line 86 "lex_sql.l" -RETURN_TOKEN(HELP); - YY_BREAK -case 9: -YY_RULE_SETUP + RETURN_TOKEN(HELP); + YY_BREAK + case 9: YY_RULE_SETUP #line 87 "lex_sql.l" -RETURN_TOKEN(DESC); - YY_BREAK -case 10: -YY_RULE_SETUP + RETURN_TOKEN(DESC); + YY_BREAK + case 10: YY_RULE_SETUP #line 88 "lex_sql.l" -RETURN_TOKEN(CREATE); - YY_BREAK -case 11: -YY_RULE_SETUP + RETURN_TOKEN(CREATE); + YY_BREAK + case 11: YY_RULE_SETUP #line 89 "lex_sql.l" -RETURN_TOKEN(DROP); - YY_BREAK -case 12: -YY_RULE_SETUP + RETURN_TOKEN(DROP); + YY_BREAK + case 12: YY_RULE_SETUP #line 90 "lex_sql.l" -RETURN_TOKEN(TABLE); - YY_BREAK -case 13: -YY_RULE_SETUP + RETURN_TOKEN(TABLE); + YY_BREAK + case 13: YY_RULE_SETUP #line 91 "lex_sql.l" -RETURN_TOKEN(TABLES); - YY_BREAK -case 14: -YY_RULE_SETUP + RETURN_TOKEN(TABLES); + YY_BREAK + case 14: YY_RULE_SETUP #line 92 "lex_sql.l" -RETURN_TOKEN(INDEX); - YY_BREAK -case 15: -YY_RULE_SETUP + RETURN_TOKEN(INDEX); + YY_BREAK + case 15: YY_RULE_SETUP #line 93 "lex_sql.l" -RETURN_TOKEN(ON); - YY_BREAK -case 16: -YY_RULE_SETUP + RETURN_TOKEN(ON); + YY_BREAK + case 16: YY_RULE_SETUP #line 94 "lex_sql.l" -RETURN_TOKEN(SHOW); - YY_BREAK -case 17: -YY_RULE_SETUP + RETURN_TOKEN(SHOW); + YY_BREAK + case 17: YY_RULE_SETUP #line 95 "lex_sql.l" -RETURN_TOKEN(SYNC); - YY_BREAK -case 18: -YY_RULE_SETUP + RETURN_TOKEN(SYNC); + YY_BREAK + case 18: YY_RULE_SETUP #line 96 "lex_sql.l" -RETURN_TOKEN(SELECT); - YY_BREAK -case 19: -YY_RULE_SETUP + RETURN_TOKEN(SELECT); + YY_BREAK + case 19: YY_RULE_SETUP #line 97 "lex_sql.l" -RETURN_TOKEN(CALC); - YY_BREAK -case 20: -YY_RULE_SETUP + RETURN_TOKEN(CALC); + YY_BREAK + case 20: YY_RULE_SETUP #line 98 "lex_sql.l" -RETURN_TOKEN(FROM); - YY_BREAK -case 21: -YY_RULE_SETUP + RETURN_TOKEN(FROM); + YY_BREAK + case 21: YY_RULE_SETUP #line 99 "lex_sql.l" -RETURN_TOKEN(WHERE); - YY_BREAK -case 22: -YY_RULE_SETUP + RETURN_TOKEN(WHERE); + YY_BREAK + case 22: YY_RULE_SETUP #line 100 "lex_sql.l" -RETURN_TOKEN(AND); - YY_BREAK -case 23: -YY_RULE_SETUP + RETURN_TOKEN(AND); + YY_BREAK + case 23: YY_RULE_SETUP #line 101 "lex_sql.l" -RETURN_TOKEN(INSERT); - YY_BREAK -case 24: -YY_RULE_SETUP + RETURN_TOKEN(OR); + YY_BREAK + case 24: YY_RULE_SETUP #line 102 "lex_sql.l" -RETURN_TOKEN(INTO); - YY_BREAK -case 25: -YY_RULE_SETUP + RETURN_TOKEN(INSERT); + YY_BREAK + case 25: YY_RULE_SETUP #line 103 "lex_sql.l" -RETURN_TOKEN(VALUES); - YY_BREAK -case 26: -YY_RULE_SETUP + RETURN_TOKEN(INTO); + YY_BREAK + case 26: YY_RULE_SETUP #line 104 "lex_sql.l" -RETURN_TOKEN(DELETE); - YY_BREAK -case 27: -YY_RULE_SETUP + RETURN_TOKEN(VALUES); + YY_BREAK + case 27: YY_RULE_SETUP #line 105 "lex_sql.l" -RETURN_TOKEN(UPDATE); - YY_BREAK -case 28: -YY_RULE_SETUP + RETURN_TOKEN(DELETE); + YY_BREAK + case 28: YY_RULE_SETUP #line 106 "lex_sql.l" -RETURN_TOKEN(SET); - YY_BREAK -case 29: -YY_RULE_SETUP + RETURN_TOKEN(UPDATE); + YY_BREAK + case 29: YY_RULE_SETUP #line 107 "lex_sql.l" -RETURN_TOKEN(TRX_BEGIN); - YY_BREAK -case 30: -YY_RULE_SETUP + RETURN_TOKEN(SET); + YY_BREAK + case 30: YY_RULE_SETUP #line 108 "lex_sql.l" -RETURN_TOKEN(TRX_COMMIT); - YY_BREAK -case 31: -YY_RULE_SETUP + RETURN_TOKEN(TRX_BEGIN); + YY_BREAK + case 31: YY_RULE_SETUP #line 109 "lex_sql.l" -RETURN_TOKEN(TRX_ROLLBACK); - YY_BREAK -case 32: -YY_RULE_SETUP + RETURN_TOKEN(TRX_COMMIT); + YY_BREAK + case 32: YY_RULE_SETUP #line 110 "lex_sql.l" -RETURN_TOKEN(INT_T); - YY_BREAK -case 33: -YY_RULE_SETUP + RETURN_TOKEN(TRX_ROLLBACK); + YY_BREAK + case 33: YY_RULE_SETUP #line 111 "lex_sql.l" -RETURN_TOKEN(STRING_T); - YY_BREAK -case 34: -YY_RULE_SETUP + RETURN_TOKEN(INT_T); + YY_BREAK + case 34: YY_RULE_SETUP #line 112 "lex_sql.l" -RETURN_TOKEN(FLOAT_T); - YY_BREAK -case 35: -YY_RULE_SETUP + RETURN_TOKEN(STRING_T); + YY_BREAK + case 35: YY_RULE_SETUP #line 113 "lex_sql.l" -RETURN_TOKEN(VECTOR_T); - YY_BREAK -case 36: -YY_RULE_SETUP + RETURN_TOKEN(FLOAT_T); + YY_BREAK + case 36: YY_RULE_SETUP #line 114 "lex_sql.l" -RETURN_TOKEN(LOAD); - YY_BREAK -case 37: -YY_RULE_SETUP + RETURN_TOKEN(DATE_T); + YY_BREAK + case 37: YY_RULE_SETUP #line 115 "lex_sql.l" -RETURN_TOKEN(DATA); - YY_BREAK -case 38: -YY_RULE_SETUP + RETURN_TOKEN(TEXT_T); + YY_BREAK + case 38: YY_RULE_SETUP #line 116 "lex_sql.l" -RETURN_TOKEN(INFILE); - YY_BREAK -case 39: -YY_RULE_SETUP + RETURN_TOKEN(VECTOR_T); + YY_BREAK + case 39: YY_RULE_SETUP #line 117 "lex_sql.l" -RETURN_TOKEN(EXPLAIN); - YY_BREAK -case 40: -YY_RULE_SETUP + RETURN_TOKEN(UNIQUE); + YY_BREAK + case 40: YY_RULE_SETUP #line 118 "lex_sql.l" -RETURN_TOKEN(GROUP); - YY_BREAK -case 41: -YY_RULE_SETUP + RETURN_TOKEN(NULL_T); + YY_BREAK + case 41: YY_RULE_SETUP #line 119 "lex_sql.l" -RETURN_TOKEN(BY); - YY_BREAK -case 42: -YY_RULE_SETUP + RETURN_TOKEN(NULLABLE); + YY_BREAK + case 42: YY_RULE_SETUP #line 120 "lex_sql.l" -RETURN_TOKEN(STORAGE); - YY_BREAK -case 43: -YY_RULE_SETUP + RETURN_TOKEN(INFILE); + YY_BREAK + case 43: YY_RULE_SETUP #line 121 "lex_sql.l" -RETURN_TOKEN(FORMAT); - YY_BREAK -case 44: -YY_RULE_SETUP + RETURN_TOKEN(EXPLAIN); + YY_BREAK + case 44: YY_RULE_SETUP #line 122 "lex_sql.l" -yylval->string=strdup(yytext); RETURN_TOKEN(ID); - YY_BREAK -case 45: -YY_RULE_SETUP + RETURN_TOKEN(GROUP); + YY_BREAK + case 45: YY_RULE_SETUP #line 123 "lex_sql.l" -RETURN_TOKEN(LBRACE); - YY_BREAK -case 46: -YY_RULE_SETUP + RETURN_TOKEN(LIMIT); + YY_BREAK + case 46: YY_RULE_SETUP #line 124 "lex_sql.l" -RETURN_TOKEN(RBRACE); - YY_BREAK -case 47: -YY_RULE_SETUP + RETURN_TOKEN(HAVING); + YY_BREAK + case 47: YY_RULE_SETUP +#line 125 "lex_sql.l" + RETURN_TOKEN(ORDER); + YY_BREAK + case 48: YY_RULE_SETUP #line 126 "lex_sql.l" -RETURN_TOKEN(COMMA); - YY_BREAK -case 48: -YY_RULE_SETUP + RETURN_TOKEN(BY); + YY_BREAK + case 49: YY_RULE_SETUP #line 127 "lex_sql.l" -RETURN_TOKEN(EQ); - YY_BREAK -case 49: -YY_RULE_SETUP + RETURN_TOKEN(AS); + YY_BREAK + case 50: YY_RULE_SETUP #line 128 "lex_sql.l" -RETURN_TOKEN(LE); - YY_BREAK -case 50: -YY_RULE_SETUP + RETURN_TOKEN(ASC); + YY_BREAK + case 51: YY_RULE_SETUP #line 129 "lex_sql.l" -RETURN_TOKEN(NE); - YY_BREAK -case 51: -YY_RULE_SETUP + RETURN_TOKEN(IN); + YY_BREAK + case 52: YY_RULE_SETUP #line 130 "lex_sql.l" -RETURN_TOKEN(NE); - YY_BREAK -case 52: -YY_RULE_SETUP + RETURN_TOKEN(EXISTS); + YY_BREAK + case 53: YY_RULE_SETUP #line 131 "lex_sql.l" -RETURN_TOKEN(LT); - YY_BREAK -case 53: -YY_RULE_SETUP + RETURN_TOKEN(STORAGE); + YY_BREAK + case 54: YY_RULE_SETUP #line 132 "lex_sql.l" -RETURN_TOKEN(GE); - YY_BREAK -case 54: -YY_RULE_SETUP + RETURN_TOKEN(FORMAT); + YY_BREAK + case 55: YY_RULE_SETUP #line 133 "lex_sql.l" -RETURN_TOKEN(GT); - YY_BREAK -case 55: + RETURN_TOKEN(INNER); + YY_BREAK + case 56: YY_RULE_SETUP +#line 134 "lex_sql.l" + RETURN_TOKEN(JOIN); + YY_BREAK + case 57: YY_RULE_SETUP +#line 135 "lex_sql.l" + RETURN_TOKEN(VIEW); + YY_BREAK + case 58: YY_RULE_SETUP #line 136 "lex_sql.l" -case 56: + RETURN_TOKEN(WITH); + YY_BREAK + case 59: YY_RULE_SETUP #line 137 "lex_sql.l" -case 57: + RETURN_TOKEN(DISTANCE); + YY_BREAK + case 60: YY_RULE_SETUP #line 138 "lex_sql.l" -case 58: -YY_RULE_SETUP -#line 138 "lex_sql.l" -{ return yytext[0]; } - YY_BREAK -case 59: -/* rule 59 can match eol */ -YY_RULE_SETUP + RETURN_TOKEN(TYPE); + YY_BREAK + case 61: YY_RULE_SETUP #line 139 "lex_sql.l" -yylval->string = strdup(yytext); RETURN_TOKEN(SSS); - YY_BREAK -case 60: -/* rule 60 can match eol */ -YY_RULE_SETUP + RETURN_TOKEN(LISTS); + YY_BREAK + case 62: YY_RULE_SETUP #line 140 "lex_sql.l" -yylval->string = strdup(yytext); RETURN_TOKEN(SSS); - YY_BREAK -case 61: -YY_RULE_SETUP + RETURN_TOKEN(PROBES); + YY_BREAK + case 63: YY_RULE_SETUP +#line 141 "lex_sql.l" + RETURN_TOKEN(IVFFLAT); + YY_BREAK + case 64: YY_RULE_SETUP #line 142 "lex_sql.l" -LOG_DEBUG("Unknown character [%c]",yytext[0]); return yytext[0]; - YY_BREAK -case 62: -YY_RULE_SETUP + RETURN_TOKEN(TRUE); + YY_BREAK + case 65: YY_RULE_SETUP #line 143 "lex_sql.l" -ECHO; - YY_BREAK -#line 1313 "lex_sql.cpp" -case YY_STATE_EOF(INITIAL): -case YY_STATE_EOF(STR): - yyterminate(); - - case YY_END_OF_BUFFER: - { - /* Amount of text matched not including the EOB char. */ - int yy_amount_of_matched_text = (int) (yy_cp - yyg->yytext_ptr) - 1; - - /* Undo the effects of YY_DO_BEFORE_ACTION. */ - *yy_cp = yyg->yy_hold_char; - YY_RESTORE_YY_MORE_OFFSET - - if ( YY_CURRENT_BUFFER_LVALUE->yy_buffer_status == YY_BUFFER_NEW ) - { - /* We're scanning a new file or input source. It's - * possible that this happened because the user - * just pointed yyin at a new source and called - * yylex(). If so, then we have to assure - * consistency between YY_CURRENT_BUFFER and our - * globals. Here is the right place to do so, because - * this is the first action (other than possibly a - * back-up) that will match for the new input source. - */ - yyg->yy_n_chars = YY_CURRENT_BUFFER_LVALUE->yy_n_chars; - YY_CURRENT_BUFFER_LVALUE->yy_input_file = yyin; - YY_CURRENT_BUFFER_LVALUE->yy_buffer_status = YY_BUFFER_NORMAL; - } - - /* Note that here we test for yy_c_buf_p "<=" to the position - * of the first EOB in the buffer, since yy_c_buf_p will - * already have been incremented past the NUL character - * (since all states make transitions on EOB to the - * end-of-buffer state). Contrast this with the test - * in input(). - */ - if ( yyg->yy_c_buf_p <= &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars] ) - { /* This was really a NUL. */ - yy_state_type yy_next_state; - - yyg->yy_c_buf_p = yyg->yytext_ptr + yy_amount_of_matched_text; - - yy_current_state = yy_get_previous_state( yyscanner ); - - /* Okay, we're now positioned to make the NUL - * transition. We couldn't have - * yy_get_previous_state() go ahead and do it - * for us because it doesn't know how to deal - * with the possibility of jamming (and we don't - * want to build jamming into it because then it - * will run more slowly). - */ - - yy_next_state = yy_try_NUL_trans( yy_current_state , yyscanner); - - yy_bp = yyg->yytext_ptr + YY_MORE_ADJ; - - if ( yy_next_state ) - { - /* Consume the NUL. */ - yy_cp = ++yyg->yy_c_buf_p; - yy_current_state = yy_next_state; - goto yy_match; - } - - else - { - yy_cp = yyg->yy_c_buf_p; - goto yy_find_action; - } - } - - else switch ( yy_get_next_buffer( yyscanner ) ) - { - case EOB_ACT_END_OF_FILE: - { - yyg->yy_did_buffer_switch_on_eof = 0; - - if ( yywrap(yyscanner ) ) - { - /* Note: because we've taken care in - * yy_get_next_buffer() to have set up - * yytext, we can now set up - * yy_c_buf_p so that if some total - * hoser (like flex itself) wants to - * call the scanner after we return the - * YY_NULL, it'll still work - another - * YY_NULL will get returned. - */ - yyg->yy_c_buf_p = yyg->yytext_ptr + YY_MORE_ADJ; - - yy_act = YY_STATE_EOF(YY_START); - goto do_action; - } - - else - { - if ( ! yyg->yy_did_buffer_switch_on_eof ) - YY_NEW_FILE; - } - break; - } - - case EOB_ACT_CONTINUE_SCAN: - yyg->yy_c_buf_p = - yyg->yytext_ptr + yy_amount_of_matched_text; - - yy_current_state = yy_get_previous_state( yyscanner ); - - yy_cp = yyg->yy_c_buf_p; - yy_bp = yyg->yytext_ptr + YY_MORE_ADJ; - goto yy_match; - - case EOB_ACT_LAST_MATCH: - yyg->yy_c_buf_p = - &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars]; - - yy_current_state = yy_get_previous_state( yyscanner ); - - yy_cp = yyg->yy_c_buf_p; - yy_bp = yyg->yytext_ptr + YY_MORE_ADJ; - goto yy_find_action; - } - break; - } - - default: - YY_FATAL_ERROR( - "fatal flex scanner internal error--no action found" ); - } /* end of action switch */ - } /* end of scanning one token */ + RETURN_TOKEN(FALSE); + YY_BREAK + case 66: YY_RULE_SETUP +#line 144 "lex_sql.l" + RETURN_TOKEN(LBRACE); + YY_BREAK + case 67: YY_RULE_SETUP +#line 145 "lex_sql.l" + RETURN_TOKEN(RBRACE); + YY_BREAK + case 68: YY_RULE_SETUP +#line 146 "lex_sql.l" + RETURN_TOKEN(LSBRACE); + YY_BREAK + case 69: YY_RULE_SETUP +#line 147 "lex_sql.l" + RETURN_TOKEN(RSBRACE); + YY_BREAK + case 70: YY_RULE_SETUP +#line 149 "lex_sql.l" + RETURN_TOKEN(COMMA); + YY_BREAK + case 71: YY_RULE_SETUP +#line 150 "lex_sql.l" + RETURN_TOKEN(EQ); + YY_BREAK + case 72: YY_RULE_SETUP +#line 151 "lex_sql.l" + RETURN_TOKEN(LE); + YY_BREAK + case 73: YY_RULE_SETUP +#line 152 "lex_sql.l" + RETURN_TOKEN(NE); + YY_BREAK + case 74: YY_RULE_SETUP +#line 153 "lex_sql.l" + RETURN_TOKEN(NE); + YY_BREAK + case 75: YY_RULE_SETUP +#line 154 "lex_sql.l" + RETURN_TOKEN(LT); + YY_BREAK + case 76: YY_RULE_SETUP +#line 155 "lex_sql.l" + RETURN_TOKEN(GE); + YY_BREAK + case 77: YY_RULE_SETUP +#line 156 "lex_sql.l" + RETURN_TOKEN(GT); + YY_BREAK + case 78: YY_RULE_SETUP +#line 157 "lex_sql.l" + RETURN_TOKEN(NOT); + YY_BREAK + case 79: YY_RULE_SETUP +#line 158 "lex_sql.l" + RETURN_TOKEN(IS); + YY_BREAK + case 80: YY_RULE_SETUP +#line 159 "lex_sql.l" + RETURN_TOKEN(LIKE); + YY_BREAK + case 81: YY_RULE_SETUP +#line 161 "lex_sql.l" + yylval->string = strdup(yytext); + RETURN_TOKEN(ID); + YY_BREAK + case 82: +#line 164 "lex_sql.l" + case 83: +#line 165 "lex_sql.l" + case 84: +#line 166 "lex_sql.l" + case 85: YY_RULE_SETUP +#line 166 "lex_sql.l" + { + return yytext[0]; + } + YY_BREAK + case 86: + /* rule 86 can match eol */ + YY_RULE_SETUP +#line 167 "lex_sql.l" + yylval->string = strdup(yytext); + RETURN_TOKEN(SSS); + YY_BREAK + case 87: + /* rule 87 can match eol */ + YY_RULE_SETUP +#line 168 "lex_sql.l" + yylval->string = strdup(yytext); + RETURN_TOKEN(SSS); + YY_BREAK + case 88: YY_RULE_SETUP +#line 170 "lex_sql.l" + LOG_DEBUG("Unknown character [%c]",yytext[0]); + return yytext[0]; + YY_BREAK + case 89: YY_RULE_SETUP +#line 171 "lex_sql.l" + ECHO; + YY_BREAK +#line 1554 "lex_sql.cpp" + case YY_STATE_EOF(INITIAL): + case YY_STATE_EOF(STR): yyterminate(); + + case YY_END_OF_BUFFER: { + /* Amount of text matched not including the EOB char. */ + int yy_amount_of_matched_text = (int)(yy_cp - yyg->yytext_ptr) - 1; + + /* Undo the effects of YY_DO_BEFORE_ACTION. */ + *yy_cp = yyg->yy_hold_char; + YY_RESTORE_YY_MORE_OFFSET + + if (YY_CURRENT_BUFFER_LVALUE->yy_buffer_status == YY_BUFFER_NEW) { + /* We're scanning a new file or input source. It's + * possible that this happened because the user + * just pointed yyin at a new source and called + * yylex(). If so, then we have to assure + * consistency between YY_CURRENT_BUFFER and our + * globals. Here is the right place to do so, because + * this is the first action (other than possibly a + * back-up) that will match for the new input source. + */ + yyg->yy_n_chars = YY_CURRENT_BUFFER_LVALUE->yy_n_chars; + YY_CURRENT_BUFFER_LVALUE->yy_input_file = yyin; + YY_CURRENT_BUFFER_LVALUE->yy_buffer_status = YY_BUFFER_NORMAL; + } + + /* Note that here we test for yy_c_buf_p "<=" to the position + * of the first EOB in the buffer, since yy_c_buf_p will + * already have been incremented past the NUL character + * (since all states make transitions on EOB to the + * end-of-buffer state). Contrast this with the test + * in input(). + */ + if (yyg->yy_c_buf_p <= &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars]) { /* This was really a NUL. */ + yy_state_type yy_next_state; + + yyg->yy_c_buf_p = yyg->yytext_ptr + yy_amount_of_matched_text; + + yy_current_state = yy_get_previous_state(yyscanner); + + /* Okay, we're now positioned to make the NUL + * transition. We couldn't have + * yy_get_previous_state() go ahead and do it + * for us because it doesn't know how to deal + * with the possibility of jamming (and we don't + * want to build jamming into it because then it + * will run more slowly). + */ + + yy_next_state = yy_try_NUL_trans(yy_current_state, yyscanner); + + yy_bp = yyg->yytext_ptr + YY_MORE_ADJ; + + if (yy_next_state) { + /* Consume the NUL. */ + yy_cp = ++yyg->yy_c_buf_p; + yy_current_state = yy_next_state; + goto yy_match; + } + + else { + yy_cp = yyg->yy_c_buf_p; + goto yy_find_action; + } + } + + else + switch (yy_get_next_buffer(yyscanner)) { + case EOB_ACT_END_OF_FILE: { + yyg->yy_did_buffer_switch_on_eof = 0; + + if (yywrap(yyscanner)) { + /* Note: because we've taken care in + * yy_get_next_buffer() to have set up + * yytext, we can now set up + * yy_c_buf_p so that if some total + * hoser (like flex itself) wants to + * call the scanner after we return the + * YY_NULL, it'll still work - another + * YY_NULL will get returned. + */ + yyg->yy_c_buf_p = yyg->yytext_ptr + YY_MORE_ADJ; + + yy_act = YY_STATE_EOF(YY_START); + goto do_action; + } + + else { + if (!yyg->yy_did_buffer_switch_on_eof) + YY_NEW_FILE; + } + break; + } + + case EOB_ACT_CONTINUE_SCAN: + yyg->yy_c_buf_p = yyg->yytext_ptr + yy_amount_of_matched_text; + + yy_current_state = yy_get_previous_state(yyscanner); + + yy_cp = yyg->yy_c_buf_p; + yy_bp = yyg->yytext_ptr + YY_MORE_ADJ; + goto yy_match; + + case EOB_ACT_LAST_MATCH: + yyg->yy_c_buf_p = &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars]; + + yy_current_state = yy_get_previous_state(yyscanner); + + yy_cp = yyg->yy_c_buf_p; + yy_bp = yyg->yytext_ptr + YY_MORE_ADJ; + goto yy_find_action; + } + break; + } + + default: YY_FATAL_ERROR("fatal flex scanner internal error--no action found"); + } /* end of action switch */ + } /* end of scanning one token */ + } /* end of user's declarations */ } /* end of yylex */ /* yy_get_next_buffer - try to read in a new buffer @@ -1450,167 +4062,149 @@ case YY_STATE_EOF(STR): * EOB_ACT_CONTINUE_SCAN - continue scanning from current position * EOB_ACT_END_OF_FILE - end of file */ -static int yy_get_next_buffer (yyscan_t yyscanner) +static int yy_get_next_buffer(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - register char *dest = YY_CURRENT_BUFFER_LVALUE->yy_ch_buf; - register char *source = yyg->yytext_ptr; - register int number_to_move, i; - int ret_val; - - if ( yyg->yy_c_buf_p > &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars + 1] ) - YY_FATAL_ERROR( - "fatal flex scanner internal error--end of buffer missed" ); - - if ( YY_CURRENT_BUFFER_LVALUE->yy_fill_buffer == 0 ) - { /* Don't try to fill the buffer, so this is an EOF. */ - if ( yyg->yy_c_buf_p - yyg->yytext_ptr - YY_MORE_ADJ == 1 ) - { - /* We matched a single character, the EOB, so - * treat this as a final EOF. - */ - return EOB_ACT_END_OF_FILE; - } - - else - { - /* We matched some text prior to the EOB, first - * process it. - */ - return EOB_ACT_LAST_MATCH; - } - } - - /* Try to read more data. */ - - /* First move last chars to start of buffer. */ - number_to_move = (int) (yyg->yy_c_buf_p - yyg->yytext_ptr) - 1; - - for ( i = 0; i < number_to_move; ++i ) - *(dest++) = *(source++); - - if ( YY_CURRENT_BUFFER_LVALUE->yy_buffer_status == YY_BUFFER_EOF_PENDING ) - /* don't do the read, it's not guaranteed to return an EOF, - * just force an EOF - */ - YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars = 0; - - else - { - int num_to_read = - YY_CURRENT_BUFFER_LVALUE->yy_buf_size - number_to_move - 1; - - while ( num_to_read <= 0 ) - { /* Not enough room in the buffer - grow it. */ - - /* just a shorter name for the current buffer */ - YY_BUFFER_STATE b = YY_CURRENT_BUFFER; - - int yy_c_buf_p_offset = - (int) (yyg->yy_c_buf_p - b->yy_ch_buf); - - if ( b->yy_is_our_buffer ) - { - int new_size = b->yy_buf_size * 2; - - if ( new_size <= 0 ) - b->yy_buf_size += b->yy_buf_size / 8; - else - b->yy_buf_size *= 2; - - b->yy_ch_buf = (char *) - /* Include room in for 2 EOB chars. */ - yyrealloc((void *) b->yy_ch_buf,b->yy_buf_size + 2 ,yyscanner ); - } - else - /* Can't grow it, we don't own it. */ - b->yy_ch_buf = 0; - - if ( ! b->yy_ch_buf ) - YY_FATAL_ERROR( - "fatal error - scanner input buffer overflow" ); - - yyg->yy_c_buf_p = &b->yy_ch_buf[yy_c_buf_p_offset]; - - num_to_read = YY_CURRENT_BUFFER_LVALUE->yy_buf_size - - number_to_move - 1; - - } - - if ( num_to_read > YY_READ_BUF_SIZE ) - num_to_read = YY_READ_BUF_SIZE; - - /* Read in more data. */ - YY_INPUT( (&YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[number_to_move]), - yyg->yy_n_chars, (size_t) num_to_read ); - - YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars; - } - - if ( yyg->yy_n_chars == 0 ) - { - if ( number_to_move == YY_MORE_ADJ ) - { - ret_val = EOB_ACT_END_OF_FILE; - yyrestart(yyin ,yyscanner); - } - - else - { - ret_val = EOB_ACT_LAST_MATCH; - YY_CURRENT_BUFFER_LVALUE->yy_buffer_status = - YY_BUFFER_EOF_PENDING; - } - } - - else - ret_val = EOB_ACT_CONTINUE_SCAN; - - if ((yy_size_t) (yyg->yy_n_chars + number_to_move) > YY_CURRENT_BUFFER_LVALUE->yy_buf_size) { - /* Extend the array by 50%, plus the number we really need. */ - yy_size_t new_size = yyg->yy_n_chars + number_to_move + (yyg->yy_n_chars >> 1); - YY_CURRENT_BUFFER_LVALUE->yy_ch_buf = (char *) yyrealloc((void *) YY_CURRENT_BUFFER_LVALUE->yy_ch_buf,new_size ,yyscanner ); - if ( ! YY_CURRENT_BUFFER_LVALUE->yy_ch_buf ) - YY_FATAL_ERROR( "out of dynamic memory in yy_get_next_buffer()" ); - } - - yyg->yy_n_chars += number_to_move; - YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars] = YY_END_OF_BUFFER_CHAR; - YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars + 1] = YY_END_OF_BUFFER_CHAR; - - yyg->yytext_ptr = &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[0]; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + char *dest = YY_CURRENT_BUFFER_LVALUE->yy_ch_buf; + char *source = yyg->yytext_ptr; + int number_to_move, i; + int ret_val; + + if (yyg->yy_c_buf_p > &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars + 1]) + YY_FATAL_ERROR("fatal flex scanner internal error--end of buffer missed"); + + if (YY_CURRENT_BUFFER_LVALUE->yy_fill_buffer == 0) { /* Don't try to fill the buffer, so this is an EOF. */ + if (yyg->yy_c_buf_p - yyg->yytext_ptr - YY_MORE_ADJ == 1) { + /* We matched a single character, the EOB, so + * treat this as a final EOF. + */ + return EOB_ACT_END_OF_FILE; + } + + else { + /* We matched some text prior to the EOB, first + * process it. + */ + return EOB_ACT_LAST_MATCH; + } + } + + /* Try to read more data. */ + + /* First move last chars to start of buffer. */ + number_to_move = (int)(yyg->yy_c_buf_p - yyg->yytext_ptr - 1); + + for (i = 0; i < number_to_move; ++i) + *(dest++) = *(source++); + + if (YY_CURRENT_BUFFER_LVALUE->yy_buffer_status == YY_BUFFER_EOF_PENDING) + /* don't do the read, it's not guaranteed to return an EOF, + * just force an EOF + */ + YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars = 0; + + else { + yy_size_t num_to_read = YY_CURRENT_BUFFER_LVALUE->yy_buf_size - number_to_move - 1; + + while (num_to_read <= 0) { /* Not enough room in the buffer - grow it. */ + + /* just a shorter name for the current buffer */ + YY_BUFFER_STATE b = YY_CURRENT_BUFFER_LVALUE; + + int yy_c_buf_p_offset = (int)(yyg->yy_c_buf_p - b->yy_ch_buf); + + if (b->yy_is_our_buffer) { + yy_size_t new_size = b->yy_buf_size * 2; + + if (new_size <= 0) + b->yy_buf_size += b->yy_buf_size / 8; + else + b->yy_buf_size *= 2; + + b->yy_ch_buf = (char *) + /* Include room in for 2 EOB chars. */ + yyrealloc((void *)b->yy_ch_buf, (yy_size_t)(b->yy_buf_size + 2), yyscanner); + } else + /* Can't grow it, we don't own it. */ + b->yy_ch_buf = NULL; + + if (!b->yy_ch_buf) + YY_FATAL_ERROR("fatal error - scanner input buffer overflow"); - return ret_val; + yyg->yy_c_buf_p = &b->yy_ch_buf[yy_c_buf_p_offset]; + + num_to_read = YY_CURRENT_BUFFER_LVALUE->yy_buf_size - number_to_move - 1; + } + + if (num_to_read > YY_READ_BUF_SIZE) + num_to_read = YY_READ_BUF_SIZE; + + /* Read in more data. */ + YY_INPUT((&YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[number_to_move]), yyg->yy_n_chars, num_to_read); + + YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars; + } + + if (yyg->yy_n_chars == 0) { + if (number_to_move == YY_MORE_ADJ) { + ret_val = EOB_ACT_END_OF_FILE; + yyrestart(yyin, yyscanner); + } + + else { + ret_val = EOB_ACT_LAST_MATCH; + YY_CURRENT_BUFFER_LVALUE->yy_buffer_status = YY_BUFFER_EOF_PENDING; + } + } + + else + ret_val = EOB_ACT_CONTINUE_SCAN; + + if ((yyg->yy_n_chars + number_to_move) > YY_CURRENT_BUFFER_LVALUE->yy_buf_size) { + /* Extend the array by 50%, plus the number we really need. */ + yy_size_t new_size = yyg->yy_n_chars + number_to_move + (yyg->yy_n_chars >> 1); + YY_CURRENT_BUFFER_LVALUE->yy_ch_buf = + (char *)yyrealloc((void *)YY_CURRENT_BUFFER_LVALUE->yy_ch_buf, (yy_size_t)new_size, yyscanner); + if (!YY_CURRENT_BUFFER_LVALUE->yy_ch_buf) + YY_FATAL_ERROR("out of dynamic memory in yy_get_next_buffer()"); + /* "- 2" to take care of EOB's */ + YY_CURRENT_BUFFER_LVALUE->yy_buf_size = (int)(new_size - 2); + } + + yyg->yy_n_chars += number_to_move; + YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars] = YY_END_OF_BUFFER_CHAR; + YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars + 1] = YY_END_OF_BUFFER_CHAR; + + yyg->yytext_ptr = &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[0]; + + return ret_val; } /* yy_get_previous_state - get the state just before the EOB char was reached */ - static yy_state_type yy_get_previous_state (yyscan_t yyscanner) +static yy_state_type yy_get_previous_state(yyscan_t yyscanner) { - register yy_state_type yy_current_state; - register char *yy_cp; - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - - yy_current_state = yyg->yy_start; - - for ( yy_cp = yyg->yytext_ptr + YY_MORE_ADJ; yy_cp < yyg->yy_c_buf_p; ++yy_cp ) - { - register YY_CHAR yy_c = (*yy_cp ? yy_ec[YY_SC_TO_UI(*yy_cp)] : 1); - if ( yy_accept[yy_current_state] ) - { - yyg->yy_last_accepting_state = yy_current_state; - yyg->yy_last_accepting_cpos = yy_cp; - } - while ( yy_chk[yy_base[yy_current_state] + yy_c] != yy_current_state ) - { - yy_current_state = (int) yy_def[yy_current_state]; - if ( yy_current_state >= 188 ) - yy_c = yy_meta[(unsigned int) yy_c]; - } - yy_current_state = yy_nxt[yy_base[yy_current_state] + (unsigned int) yy_c]; - } - - return yy_current_state; + yy_state_type yy_current_state; + char *yy_cp; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + + yy_current_state = yyg->yy_start; + + for (yy_cp = yyg->yytext_ptr + YY_MORE_ADJ; yy_cp < yyg->yy_c_buf_p; ++yy_cp) { + YY_CHAR yy_c = (*yy_cp ? yy_ec[YY_SC_TO_UI(*yy_cp)] : 1); + if (yy_accept[yy_current_state]) { + yyg->yy_last_accepting_state = yy_current_state; + yyg->yy_last_accepting_cpos = yy_cp; + } + while (yy_chk[yy_base[yy_current_state] + yy_c] != yy_current_state) { + yy_current_state = (int)yy_def[yy_current_state]; + if (yy_current_state >= 271) + yy_c = yy_meta[yy_c]; + } + yy_current_state = yy_nxt[yy_base[yy_current_state] + yy_c]; + } + + return yy_current_state; } /* yy_try_NUL_trans - try to make a transition on the NUL character @@ -1618,167 +4212,162 @@ static int yy_get_next_buffer (yyscan_t yyscanner) * synopsis * next_state = yy_try_NUL_trans( current_state ); */ - static yy_state_type yy_try_NUL_trans (yy_state_type yy_current_state , yyscan_t yyscanner) +static yy_state_type yy_try_NUL_trans(yy_state_type yy_current_state, yyscan_t yyscanner) { - register int yy_is_jam; - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; /* This var may be unused depending upon options. */ - register char *yy_cp = yyg->yy_c_buf_p; - - register YY_CHAR yy_c = 1; - if ( yy_accept[yy_current_state] ) - { - yyg->yy_last_accepting_state = yy_current_state; - yyg->yy_last_accepting_cpos = yy_cp; - } - while ( yy_chk[yy_base[yy_current_state] + yy_c] != yy_current_state ) - { - yy_current_state = (int) yy_def[yy_current_state]; - if ( yy_current_state >= 188 ) - yy_c = yy_meta[(unsigned int) yy_c]; - } - yy_current_state = yy_nxt[yy_base[yy_current_state] + (unsigned int) yy_c]; - yy_is_jam = (yy_current_state == 187); - - return yy_is_jam ? 0 : yy_current_state; + int yy_is_jam; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; /* This var may be unused depending upon options. */ + char *yy_cp = yyg->yy_c_buf_p; + + YY_CHAR yy_c = 1; + if (yy_accept[yy_current_state]) { + yyg->yy_last_accepting_state = yy_current_state; + yyg->yy_last_accepting_cpos = yy_cp; + } + while (yy_chk[yy_base[yy_current_state] + yy_c] != yy_current_state) { + yy_current_state = (int)yy_def[yy_current_state]; + if (yy_current_state >= 271) + yy_c = yy_meta[yy_c]; + } + yy_current_state = yy_nxt[yy_base[yy_current_state] + yy_c]; + yy_is_jam = (yy_current_state == 270); + + (void)yyg; + return yy_is_jam ? 0 : yy_current_state; } +#ifndef YY_NO_UNPUT + +#endif + #ifndef YY_NO_INPUT #ifdef __cplusplus - static int yyinput (yyscan_t yyscanner) +static int yyinput(yyscan_t yyscanner) #else - static int input (yyscan_t yyscanner) +static int input(yyscan_t yyscanner) #endif { - int c; - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - - *yyg->yy_c_buf_p = yyg->yy_hold_char; - - if ( *yyg->yy_c_buf_p == YY_END_OF_BUFFER_CHAR ) - { - /* yy_c_buf_p now points to the character we want to return. - * If this occurs *before* the EOB characters, then it's a - * valid NUL; if not, then we've hit the end of the buffer. - */ - if ( yyg->yy_c_buf_p < &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars] ) - /* This was really a NUL. */ - *yyg->yy_c_buf_p = '\0'; - - else - { /* need more input */ - int offset = yyg->yy_c_buf_p - yyg->yytext_ptr; - ++yyg->yy_c_buf_p; - - switch ( yy_get_next_buffer( yyscanner ) ) - { - case EOB_ACT_LAST_MATCH: - /* This happens because yy_g_n_b() - * sees that we've accumulated a - * token and flags that we need to - * try matching the token before - * proceeding. But for input(), - * there's no matching to consider. - * So convert the EOB_ACT_LAST_MATCH - * to EOB_ACT_END_OF_FILE. - */ - - /* Reset buffer status. */ - yyrestart(yyin ,yyscanner); - - /*FALLTHROUGH*/ - - case EOB_ACT_END_OF_FILE: - { - if ( yywrap(yyscanner ) ) - return EOF; - - if ( ! yyg->yy_did_buffer_switch_on_eof ) - YY_NEW_FILE; + int c; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + + *yyg->yy_c_buf_p = yyg->yy_hold_char; + + if (*yyg->yy_c_buf_p == YY_END_OF_BUFFER_CHAR) { + /* yy_c_buf_p now points to the character we want to return. + * If this occurs *before* the EOB characters, then it's a + * valid NUL; if not, then we've hit the end of the buffer. + */ + if (yyg->yy_c_buf_p < &YY_CURRENT_BUFFER_LVALUE->yy_ch_buf[yyg->yy_n_chars]) + /* This was really a NUL. */ + *yyg->yy_c_buf_p = '\0'; + + else { /* need more input */ + yy_size_t offset = yyg->yy_c_buf_p - yyg->yytext_ptr; + ++yyg->yy_c_buf_p; + + switch (yy_get_next_buffer(yyscanner)) { + case EOB_ACT_LAST_MATCH: + /* This happens because yy_g_n_b() + * sees that we've accumulated a + * token and flags that we need to + * try matching the token before + * proceeding. But for input(), + * there's no matching to consider. + * So convert the EOB_ACT_LAST_MATCH + * to EOB_ACT_END_OF_FILE. + */ + + /* Reset buffer status. */ + yyrestart(yyin, yyscanner); + + /*FALLTHROUGH*/ + + case EOB_ACT_END_OF_FILE: { + if (yywrap(yyscanner)) + return 0; + + if (!yyg->yy_did_buffer_switch_on_eof) + YY_NEW_FILE; #ifdef __cplusplus - return yyinput(yyscanner); + return yyinput(yyscanner); #else - return input(yyscanner); + return input(yyscanner); #endif - } + } - case EOB_ACT_CONTINUE_SCAN: - yyg->yy_c_buf_p = yyg->yytext_ptr + offset; - break; - } - } - } + case EOB_ACT_CONTINUE_SCAN: yyg->yy_c_buf_p = yyg->yytext_ptr + offset; break; + } + } + } - c = *(unsigned char *) yyg->yy_c_buf_p; /* cast for 8-bit char's */ - *yyg->yy_c_buf_p = '\0'; /* preserve yytext */ - yyg->yy_hold_char = *++yyg->yy_c_buf_p; + c = *(unsigned char *)yyg->yy_c_buf_p; /* cast for 8-bit char's */ + *yyg->yy_c_buf_p = '\0'; /* preserve yytext */ + yyg->yy_hold_char = *++yyg->yy_c_buf_p; - return c; + return c; } -#endif /* ifndef YY_NO_INPUT */ +#endif /* ifndef YY_NO_INPUT */ /** Immediately switch to a different input stream. * @param input_file A readable stream. * @param yyscanner The scanner object. * @note This function does not reset the start condition to @c INITIAL . */ - void yyrestart (FILE * input_file , yyscan_t yyscanner) +void yyrestart(FILE *input_file, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; - if ( ! YY_CURRENT_BUFFER ){ - yyensure_buffer_stack (yyscanner); - YY_CURRENT_BUFFER_LVALUE = - yy_create_buffer(yyin,YY_BUF_SIZE ,yyscanner); - } + if (!YY_CURRENT_BUFFER) { + yyensure_buffer_stack(yyscanner); + YY_CURRENT_BUFFER_LVALUE = yy_create_buffer(yyin, YY_BUF_SIZE, yyscanner); + } - yy_init_buffer(YY_CURRENT_BUFFER,input_file ,yyscanner); - yy_load_buffer_state(yyscanner ); + yy_init_buffer(YY_CURRENT_BUFFER, input_file, yyscanner); + yy_load_buffer_state(yyscanner); } /** Switch to a different input buffer. * @param new_buffer The new input buffer. * @param yyscanner The scanner object. */ - void yy_switch_to_buffer (YY_BUFFER_STATE new_buffer , yyscan_t yyscanner) +void yy_switch_to_buffer(YY_BUFFER_STATE new_buffer, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - - /* TODO. We should be able to replace this entire function body - * with - * yypop_buffer_state(); - * yypush_buffer_state(new_buffer); - */ - yyensure_buffer_stack (yyscanner); - if ( YY_CURRENT_BUFFER == new_buffer ) - return; - - if ( YY_CURRENT_BUFFER ) - { - /* Flush out information for old buffer. */ - *yyg->yy_c_buf_p = yyg->yy_hold_char; - YY_CURRENT_BUFFER_LVALUE->yy_buf_pos = yyg->yy_c_buf_p; - YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars; - } - - YY_CURRENT_BUFFER_LVALUE = new_buffer; - yy_load_buffer_state(yyscanner ); - - /* We don't actually know whether we did this switch during - * EOF (yywrap()) processing, but the only time this flag - * is looked at is after yywrap() is called, so it's safe - * to go ahead and always set it. - */ - yyg->yy_did_buffer_switch_on_eof = 1; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + + /* TODO. We should be able to replace this entire function body + * with + * yypop_buffer_state(); + * yypush_buffer_state(new_buffer); + */ + yyensure_buffer_stack(yyscanner); + if (YY_CURRENT_BUFFER == new_buffer) + return; + + if (YY_CURRENT_BUFFER) { + /* Flush out information for old buffer. */ + *yyg->yy_c_buf_p = yyg->yy_hold_char; + YY_CURRENT_BUFFER_LVALUE->yy_buf_pos = yyg->yy_c_buf_p; + YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars; + } + + YY_CURRENT_BUFFER_LVALUE = new_buffer; + yy_load_buffer_state(yyscanner); + + /* We don't actually know whether we did this switch during + * EOF (yywrap()) processing, but the only time this flag + * is looked at is after yywrap() is called, so it's safe + * to go ahead and always set it. + */ + yyg->yy_did_buffer_switch_on_eof = 1; } -static void yy_load_buffer_state (yyscan_t yyscanner) +static void yy_load_buffer_state(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - yyg->yy_n_chars = YY_CURRENT_BUFFER_LVALUE->yy_n_chars; - yyg->yytext_ptr = yyg->yy_c_buf_p = YY_CURRENT_BUFFER_LVALUE->yy_buf_pos; - yyin = YY_CURRENT_BUFFER_LVALUE->yy_input_file; - yyg->yy_hold_char = *yyg->yy_c_buf_p; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + yyg->yy_n_chars = YY_CURRENT_BUFFER_LVALUE->yy_n_chars; + yyg->yytext_ptr = yyg->yy_c_buf_p = YY_CURRENT_BUFFER_LVALUE->yy_buf_pos; + yyin = YY_CURRENT_BUFFER_LVALUE->yy_input_file; + yyg->yy_hold_char = *yyg->yy_c_buf_p; } /** Allocate and initialize an input buffer state. @@ -1787,109 +4376,105 @@ static void yy_load_buffer_state (yyscan_t yyscanner) * @param yyscanner The scanner object. * @return the allocated buffer state. */ - YY_BUFFER_STATE yy_create_buffer (FILE * file, int size , yyscan_t yyscanner) +YY_BUFFER_STATE yy_create_buffer(FILE *file, int size, yyscan_t yyscanner) { - YY_BUFFER_STATE b; - - b = (YY_BUFFER_STATE) yyalloc(sizeof( struct yy_buffer_state ) ,yyscanner ); - if ( ! b ) - YY_FATAL_ERROR( "out of dynamic memory in yy_create_buffer()" ); + YY_BUFFER_STATE b; - b->yy_buf_size = size; + b = (YY_BUFFER_STATE)yyalloc(sizeof(struct yy_buffer_state), yyscanner); + if (!b) + YY_FATAL_ERROR("out of dynamic memory in yy_create_buffer()"); - /* yy_ch_buf has to be 2 characters longer than the size given because - * we need to put in 2 end-of-buffer characters. - */ - b->yy_ch_buf = (char *) yyalloc(b->yy_buf_size + 2 ,yyscanner ); - if ( ! b->yy_ch_buf ) - YY_FATAL_ERROR( "out of dynamic memory in yy_create_buffer()" ); + b->yy_buf_size = size; - b->yy_is_our_buffer = 1; + /* yy_ch_buf has to be 2 characters longer than the size given because + * we need to put in 2 end-of-buffer characters. + */ + b->yy_ch_buf = (char *)yyalloc((yy_size_t)(b->yy_buf_size + 2), yyscanner); + if (!b->yy_ch_buf) + YY_FATAL_ERROR("out of dynamic memory in yy_create_buffer()"); - yy_init_buffer(b,file ,yyscanner); + b->yy_is_our_buffer = 1; - return b; + yy_init_buffer(b, file, yyscanner); + + return b; } /** Destroy the buffer. * @param b a buffer created with yy_create_buffer() * @param yyscanner The scanner object. */ - void yy_delete_buffer (YY_BUFFER_STATE b , yyscan_t yyscanner) +void yy_delete_buffer(YY_BUFFER_STATE b, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; - if ( ! b ) - return; + if (!b) + return; - if ( b == YY_CURRENT_BUFFER ) /* Not sure if we should pop here. */ - YY_CURRENT_BUFFER_LVALUE = (YY_BUFFER_STATE) 0; + if (b == YY_CURRENT_BUFFER) /* Not sure if we should pop here. */ + YY_CURRENT_BUFFER_LVALUE = (YY_BUFFER_STATE)0; - if ( b->yy_is_our_buffer ) - yyfree((void *) b->yy_ch_buf ,yyscanner ); + if (b->yy_is_our_buffer) + yyfree((void *)b->yy_ch_buf, yyscanner); - yyfree((void *) b ,yyscanner ); + yyfree((void *)b, yyscanner); } -#ifndef __cplusplus -extern int isatty (int ); -#endif /* __cplusplus */ - /* Initializes or reinitializes a buffer. * This function is sometimes called more than once on the same buffer, * such as during a yyrestart() or at EOF. */ - static void yy_init_buffer (YY_BUFFER_STATE b, FILE * file , yyscan_t yyscanner) +static void yy_init_buffer(YY_BUFFER_STATE b, FILE *file, yyscan_t yyscanner) { - int oerrno = errno; - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; + int oerrno = errno; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; - yy_flush_buffer(b ,yyscanner); + yy_flush_buffer(b, yyscanner); - b->yy_input_file = file; - b->yy_fill_buffer = 1; + b->yy_input_file = file; + b->yy_fill_buffer = 1; - /* If b is the current buffer, then yy_init_buffer was _probably_ - * called from yyrestart() or through yy_get_next_buffer. - * In that case, we don't want to reset the lineno or column. - */ - if (b != YY_CURRENT_BUFFER){ - b->yy_bs_lineno = 1; - b->yy_bs_column = 0; - } + /* If b is the current buffer, then yy_init_buffer was _probably_ + * called from yyrestart() or through yy_get_next_buffer. + * In that case, we don't want to reset the lineno or column. + */ + if (b != YY_CURRENT_BUFFER) { + b->yy_bs_lineno = 1; + b->yy_bs_column = 0; + } - b->yy_is_interactive = file ? (isatty( fileno(file) ) > 0) : 0; - - errno = oerrno; + b->yy_is_interactive = file ? (isatty(fileno(file)) > 0) : 0; + + errno = oerrno; } /** Discard all buffered characters. On the next scan, YY_INPUT will be called. * @param b the buffer state to be flushed, usually @c YY_CURRENT_BUFFER. * @param yyscanner The scanner object. */ - void yy_flush_buffer (YY_BUFFER_STATE b , yyscan_t yyscanner) +void yy_flush_buffer(YY_BUFFER_STATE b, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - if ( ! b ) - return; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + if (!b) + return; - b->yy_n_chars = 0; + b->yy_n_chars = 0; - /* We always need two end-of-buffer characters. The first causes - * a transition to the end-of-buffer state. The second causes - * a jam in that state. - */ - b->yy_ch_buf[0] = YY_END_OF_BUFFER_CHAR; - b->yy_ch_buf[1] = YY_END_OF_BUFFER_CHAR; + /* We always need two end-of-buffer characters. The first causes + * a transition to the end-of-buffer state. The second causes + * a jam in that state. + */ + b->yy_ch_buf[0] = YY_END_OF_BUFFER_CHAR; + b->yy_ch_buf[1] = YY_END_OF_BUFFER_CHAR; - b->yy_buf_pos = &b->yy_ch_buf[0]; + b->yy_buf_pos = &b->yy_ch_buf[0]; - b->yy_at_bol = 1; - b->yy_buffer_status = YY_BUFFER_NEW; + b->yy_at_bol = 1; + b->yy_buffer_status = YY_BUFFER_NEW; - if ( b == YY_CURRENT_BUFFER ) - yy_load_buffer_state(yyscanner ); + if (b == YY_CURRENT_BUFFER) + yy_load_buffer_state(yyscanner); } /** Pushes the new state onto the stack. The new state becomes @@ -1898,134 +4483,128 @@ extern int isatty (int ); * @param new_buffer The new state. * @param yyscanner The scanner object. */ -void yypush_buffer_state (YY_BUFFER_STATE new_buffer , yyscan_t yyscanner) +void yypush_buffer_state(YY_BUFFER_STATE new_buffer, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - if (new_buffer == NULL) - return; - - yyensure_buffer_stack(yyscanner); - - /* This block is copied from yy_switch_to_buffer. */ - if ( YY_CURRENT_BUFFER ) - { - /* Flush out information for old buffer. */ - *yyg->yy_c_buf_p = yyg->yy_hold_char; - YY_CURRENT_BUFFER_LVALUE->yy_buf_pos = yyg->yy_c_buf_p; - YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars; - } - - /* Only push if top exists. Otherwise, replace top. */ - if (YY_CURRENT_BUFFER) - yyg->yy_buffer_stack_top++; - YY_CURRENT_BUFFER_LVALUE = new_buffer; - - /* copied from yy_switch_to_buffer. */ - yy_load_buffer_state(yyscanner ); - yyg->yy_did_buffer_switch_on_eof = 1; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + if (new_buffer == NULL) + return; + + yyensure_buffer_stack(yyscanner); + + /* This block is copied from yy_switch_to_buffer. */ + if (YY_CURRENT_BUFFER) { + /* Flush out information for old buffer. */ + *yyg->yy_c_buf_p = yyg->yy_hold_char; + YY_CURRENT_BUFFER_LVALUE->yy_buf_pos = yyg->yy_c_buf_p; + YY_CURRENT_BUFFER_LVALUE->yy_n_chars = yyg->yy_n_chars; + } + + /* Only push if top exists. Otherwise, replace top. */ + if (YY_CURRENT_BUFFER) + yyg->yy_buffer_stack_top++; + YY_CURRENT_BUFFER_LVALUE = new_buffer; + + /* copied from yy_switch_to_buffer. */ + yy_load_buffer_state(yyscanner); + yyg->yy_did_buffer_switch_on_eof = 1; } /** Removes and deletes the top of the stack, if present. * The next element becomes the new top. * @param yyscanner The scanner object. */ -void yypop_buffer_state (yyscan_t yyscanner) +void yypop_buffer_state(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - if (!YY_CURRENT_BUFFER) - return; - - yy_delete_buffer(YY_CURRENT_BUFFER ,yyscanner); - YY_CURRENT_BUFFER_LVALUE = NULL; - if (yyg->yy_buffer_stack_top > 0) - --yyg->yy_buffer_stack_top; - - if (YY_CURRENT_BUFFER) { - yy_load_buffer_state(yyscanner ); - yyg->yy_did_buffer_switch_on_eof = 1; - } + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + if (!YY_CURRENT_BUFFER) + return; + + yy_delete_buffer(YY_CURRENT_BUFFER, yyscanner); + YY_CURRENT_BUFFER_LVALUE = NULL; + if (yyg->yy_buffer_stack_top > 0) + --yyg->yy_buffer_stack_top; + + if (YY_CURRENT_BUFFER) { + yy_load_buffer_state(yyscanner); + yyg->yy_did_buffer_switch_on_eof = 1; + } } /* Allocates the stack if it does not exist. * Guarantees space for at least one push. */ -static void yyensure_buffer_stack (yyscan_t yyscanner) +static void yyensure_buffer_stack(yyscan_t yyscanner) { - int num_to_alloc; - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - - if (!yyg->yy_buffer_stack) { - - /* First allocation is just for 2 elements, since we don't know if this - * scanner will even need a stack. We use 2 instead of 1 to avoid an - * immediate realloc on the next call. - */ - num_to_alloc = 1; - yyg->yy_buffer_stack = (struct yy_buffer_state**)yyalloc - (num_to_alloc * sizeof(struct yy_buffer_state*) - , yyscanner); - if ( ! yyg->yy_buffer_stack ) - YY_FATAL_ERROR( "out of dynamic memory in yyensure_buffer_stack()" ); - - memset(yyg->yy_buffer_stack, 0, num_to_alloc * sizeof(struct yy_buffer_state*)); - - yyg->yy_buffer_stack_max = num_to_alloc; - yyg->yy_buffer_stack_top = 0; - return; - } - - if (yyg->yy_buffer_stack_top >= (yyg->yy_buffer_stack_max) - 1){ - - /* Increase the buffer to prepare for a possible push. */ - int grow_size = 8 /* arbitrary grow size */; - - num_to_alloc = yyg->yy_buffer_stack_max + grow_size; - yyg->yy_buffer_stack = (struct yy_buffer_state**)yyrealloc - (yyg->yy_buffer_stack, - num_to_alloc * sizeof(struct yy_buffer_state*) - , yyscanner); - if ( ! yyg->yy_buffer_stack ) - YY_FATAL_ERROR( "out of dynamic memory in yyensure_buffer_stack()" ); - - /* zero only the new slots.*/ - memset(yyg->yy_buffer_stack + yyg->yy_buffer_stack_max, 0, grow_size * sizeof(struct yy_buffer_state*)); - yyg->yy_buffer_stack_max = num_to_alloc; - } + yy_size_t num_to_alloc; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + + if (!yyg->yy_buffer_stack) { + + /* First allocation is just for 2 elements, since we don't know if this + * scanner will even need a stack. We use 2 instead of 1 to avoid an + * immediate realloc on the next call. + */ + num_to_alloc = 1; /* After all that talk, this was set to 1 anyways... */ + yyg->yy_buffer_stack = + (struct yy_buffer_state **)yyalloc(num_to_alloc * sizeof(struct yy_buffer_state *), yyscanner); + if (!yyg->yy_buffer_stack) + YY_FATAL_ERROR("out of dynamic memory in yyensure_buffer_stack()"); + + memset(yyg->yy_buffer_stack, 0, num_to_alloc * sizeof(struct yy_buffer_state *)); + + yyg->yy_buffer_stack_max = num_to_alloc; + yyg->yy_buffer_stack_top = 0; + return; + } + + if (yyg->yy_buffer_stack_top >= (yyg->yy_buffer_stack_max) - 1) { + + /* Increase the buffer to prepare for a possible push. */ + yy_size_t grow_size = 8 /* arbitrary grow size */; + + num_to_alloc = yyg->yy_buffer_stack_max + grow_size; + yyg->yy_buffer_stack = (struct yy_buffer_state **)yyrealloc( + yyg->yy_buffer_stack, num_to_alloc * sizeof(struct yy_buffer_state *), yyscanner); + if (!yyg->yy_buffer_stack) + YY_FATAL_ERROR("out of dynamic memory in yyensure_buffer_stack()"); + + /* zero only the new slots.*/ + memset(yyg->yy_buffer_stack + yyg->yy_buffer_stack_max, 0, grow_size * sizeof(struct yy_buffer_state *)); + yyg->yy_buffer_stack_max = num_to_alloc; + } } /** Setup the input buffer state to scan directly from a user-specified character buffer. * @param base the character buffer * @param size the size in bytes of the character buffer * @param yyscanner The scanner object. - * @return the newly allocated buffer state object. + * @return the newly allocated buffer state object. */ -YY_BUFFER_STATE yy_scan_buffer (char * base, yy_size_t size , yyscan_t yyscanner) +YY_BUFFER_STATE yy_scan_buffer(char *base, yy_size_t size, yyscan_t yyscanner) { - YY_BUFFER_STATE b; - - if ( size < 2 || - base[size-2] != YY_END_OF_BUFFER_CHAR || - base[size-1] != YY_END_OF_BUFFER_CHAR ) - /* They forgot to leave room for the EOB's. */ - return 0; - - b = (YY_BUFFER_STATE) yyalloc(sizeof( struct yy_buffer_state ) ,yyscanner ); - if ( ! b ) - YY_FATAL_ERROR( "out of dynamic memory in yy_scan_buffer()" ); - - b->yy_buf_size = size - 2; /* "- 2" to take care of EOB's */ - b->yy_buf_pos = b->yy_ch_buf = base; - b->yy_is_our_buffer = 0; - b->yy_input_file = 0; - b->yy_n_chars = b->yy_buf_size; - b->yy_is_interactive = 0; - b->yy_at_bol = 1; - b->yy_fill_buffer = 0; - b->yy_buffer_status = YY_BUFFER_NEW; - - yy_switch_to_buffer(b ,yyscanner ); - - return b; + YY_BUFFER_STATE b; + + if (size < 2 || base[size - 2] != YY_END_OF_BUFFER_CHAR || base[size - 1] != YY_END_OF_BUFFER_CHAR) + /* They forgot to leave room for the EOB's. */ + return NULL; + + b = (YY_BUFFER_STATE)yyalloc(sizeof(struct yy_buffer_state), yyscanner); + if (!b) + YY_FATAL_ERROR("out of dynamic memory in yy_scan_buffer()"); + + b->yy_buf_size = (int)(size - 2); /* "- 2" to take care of EOB's */ + b->yy_buf_pos = b->yy_ch_buf = base; + b->yy_is_our_buffer = 0; + b->yy_input_file = NULL; + b->yy_n_chars = b->yy_buf_size; + b->yy_is_interactive = 0; + b->yy_at_bol = 1; + b->yy_fill_buffer = 0; + b->yy_buffer_status = YY_BUFFER_NEW; + + yy_switch_to_buffer(b, yyscanner); + + return b; } /** Setup the input buffer state to scan a string. The next call to yylex() will @@ -2036,272 +4615,270 @@ YY_BUFFER_STATE yy_scan_buffer (char * base, yy_size_t size , yyscan_t yyscann * @note If you want to scan bytes that may contain NUL values, then use * yy_scan_bytes() instead. */ -YY_BUFFER_STATE yy_scan_string (yyconst char * yystr , yyscan_t yyscanner) +YY_BUFFER_STATE yy_scan_string(const char *yystr, yyscan_t yyscanner) { - - return yy_scan_bytes(yystr,strlen(yystr) ,yyscanner); + + return yy_scan_bytes(yystr, (int)strlen(yystr), yyscanner); } /** Setup the input buffer state to scan the given bytes. The next call to yylex() will * scan from a @e copy of @a bytes. - * @param bytes the byte buffer to scan - * @param len the number of bytes in the buffer pointed to by @a bytes. + * @param yybytes the byte buffer to scan + * @param _yybytes_len the number of bytes in the buffer pointed to by @a bytes. * @param yyscanner The scanner object. * @return the newly allocated buffer state object. */ -YY_BUFFER_STATE yy_scan_bytes (yyconst char * yybytes, int _yybytes_len , yyscan_t yyscanner) +YY_BUFFER_STATE yy_scan_bytes(const char *yybytes, yy_size_t _yybytes_len, yyscan_t yyscanner) { - YY_BUFFER_STATE b; - char *buf; - yy_size_t n; - int i; - - /* Get memory for full buffer, including space for trailing EOB's. */ - n = _yybytes_len + 2; - buf = (char *) yyalloc(n ,yyscanner ); - if ( ! buf ) - YY_FATAL_ERROR( "out of dynamic memory in yy_scan_bytes()" ); - - for ( i = 0; i < _yybytes_len; ++i ) - buf[i] = yybytes[i]; - - buf[_yybytes_len] = buf[_yybytes_len+1] = YY_END_OF_BUFFER_CHAR; - - b = yy_scan_buffer(buf,n ,yyscanner); - if ( ! b ) - YY_FATAL_ERROR( "bad buffer in yy_scan_bytes()" ); - - /* It's okay to grow etc. this buffer, and we should throw it - * away when we're done. - */ - b->yy_is_our_buffer = 1; - - return b; + YY_BUFFER_STATE b; + char *buf; + yy_size_t n; + yy_size_t i; + + /* Get memory for full buffer, including space for trailing EOB's. */ + n = (yy_size_t)(_yybytes_len + 2); + buf = (char *)yyalloc(n, yyscanner); + if (!buf) + YY_FATAL_ERROR("out of dynamic memory in yy_scan_bytes()"); + + for (i = 0; i < _yybytes_len; ++i) + buf[i] = yybytes[i]; + + buf[_yybytes_len] = buf[_yybytes_len + 1] = YY_END_OF_BUFFER_CHAR; + + b = yy_scan_buffer(buf, n, yyscanner); + if (!b) + YY_FATAL_ERROR("bad buffer in yy_scan_bytes()"); + + /* It's okay to grow etc. this buffer, and we should throw it + * away when we're done. + */ + b->yy_is_our_buffer = 1; + + return b; } #ifndef YY_EXIT_FAILURE #define YY_EXIT_FAILURE 2 #endif -static void yy_fatal_error (yyconst char* msg , yyscan_t yyscanner) +static void yynoreturn yy_fatal_error(const char *msg, yyscan_t yyscanner) { - (void) fprintf( stderr, "%s\n", msg ); - exit( YY_EXIT_FAILURE ); + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + (void)yyg; + fprintf(stderr, "%s\n", msg); + exit(YY_EXIT_FAILURE); } /* Redefine yyless() so it works in section 3 code. */ #undef yyless -#define yyless(n) \ - do \ - { \ - /* Undo effects of setting up yytext. */ \ - int yyless_macro_arg = (n); \ - YY_LESS_LINENO(yyless_macro_arg);\ - yytext[yyleng] = yyg->yy_hold_char; \ - yyg->yy_c_buf_p = yytext + yyless_macro_arg; \ - yyg->yy_hold_char = *yyg->yy_c_buf_p; \ - *yyg->yy_c_buf_p = '\0'; \ - yyleng = yyless_macro_arg; \ - } \ - while ( 0 ) +#define yyless(n) \ + do { \ + /* Undo effects of setting up yytext. */ \ + yy_size_t yyless_macro_arg = (n); \ + YY_LESS_LINENO(yyless_macro_arg); \ + yytext[yyleng] = yyg->yy_hold_char; \ + yyg->yy_c_buf_p = yytext + yyless_macro_arg; \ + yyg->yy_hold_char = *yyg->yy_c_buf_p; \ + *yyg->yy_c_buf_p = '\0'; \ + yyleng = yyless_macro_arg; \ + } while (0) /* Accessor methods (get/set functions) to struct members. */ /** Get the user-defined data for this scanner. * @param yyscanner The scanner object. */ -YY_EXTRA_TYPE yyget_extra (yyscan_t yyscanner) +YY_EXTRA_TYPE yyget_extra(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - return yyextra; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + return yyextra; } /** Get the current line number. * @param yyscanner The scanner object. */ -int yyget_lineno (yyscan_t yyscanner) +int yyget_lineno(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - - if (! YY_CURRENT_BUFFER) - return 0; - - return yylineno; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + + if (!YY_CURRENT_BUFFER) + return 0; + + return yylineno; } /** Get the current column number. * @param yyscanner The scanner object. */ -int yyget_column (yyscan_t yyscanner) +int yyget_column(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - - if (! YY_CURRENT_BUFFER) - return 0; - - return yycolumn; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + + if (!YY_CURRENT_BUFFER) + return 0; + + return yycolumn; } /** Get the input stream. * @param yyscanner The scanner object. */ -FILE *yyget_in (yyscan_t yyscanner) +FILE *yyget_in(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - return yyin; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + return yyin; } /** Get the output stream. * @param yyscanner The scanner object. */ -FILE *yyget_out (yyscan_t yyscanner) +FILE *yyget_out(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - return yyout; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + return yyout; } /** Get the length of the current token. * @param yyscanner The scanner object. */ -int yyget_leng (yyscan_t yyscanner) +yy_size_t yyget_leng(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - return yyleng; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + return yyleng; } /** Get the current token. * @param yyscanner The scanner object. */ -char *yyget_text (yyscan_t yyscanner) +char *yyget_text(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - return yytext; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + return yytext; } /** Set the user-defined data. This data is never touched by the scanner. * @param user_defined The data to be associated with this scanner. * @param yyscanner The scanner object. */ -void yyset_extra (YY_EXTRA_TYPE user_defined , yyscan_t yyscanner) +void yyset_extra(YY_EXTRA_TYPE user_defined, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - yyextra = user_defined ; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + yyextra = user_defined; } /** Set the current line number. - * @param line_number + * @param _line_number line number * @param yyscanner The scanner object. */ -void yyset_lineno (int line_number , yyscan_t yyscanner) +void yyset_lineno(int _line_number, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; - /* lineno is only valid if an input buffer exists. */ - if (! YY_CURRENT_BUFFER ) - yy_fatal_error( "yyset_lineno called with no buffer" , yyscanner); - - yylineno = line_number; + /* lineno is only valid if an input buffer exists. */ + if (!YY_CURRENT_BUFFER) + YY_FATAL_ERROR("yyset_lineno called with no buffer"); + + yylineno = _line_number; } /** Set the current column. - * @param line_number + * @param _column_no column number * @param yyscanner The scanner object. */ -void yyset_column (int column_no , yyscan_t yyscanner) +void yyset_column(int _column_no, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + + /* column is only valid if an input buffer exists. */ + if (!YY_CURRENT_BUFFER) + YY_FATAL_ERROR("yyset_column called with no buffer"); - /* column is only valid if an input buffer exists. */ - if (! YY_CURRENT_BUFFER ) - yy_fatal_error( "yyset_column called with no buffer" , yyscanner); - - yycolumn = column_no; + yycolumn = _column_no; } /** Set the input stream. This does not discard the current * input buffer. - * @param in_str A readable stream. + * @param _in_str A readable stream. * @param yyscanner The scanner object. * @see yy_switch_to_buffer */ -void yyset_in (FILE * in_str , yyscan_t yyscanner) +void yyset_in(FILE *_in_str, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - yyin = in_str ; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + yyin = _in_str; } -void yyset_out (FILE * out_str , yyscan_t yyscanner) +void yyset_out(FILE *_out_str, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - yyout = out_str ; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + yyout = _out_str; } -int yyget_debug (yyscan_t yyscanner) +int yyget_debug(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - return yy_flex_debug; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + return yy_flex_debug; } -void yyset_debug (int bdebug , yyscan_t yyscanner) +void yyset_debug(int _bdebug, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - yy_flex_debug = bdebug ; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + yy_flex_debug = _bdebug; } /* Accessor methods for yylval and yylloc */ -YYSTYPE * yyget_lval (yyscan_t yyscanner) +YYSTYPE *yyget_lval(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - return yylval; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + return yylval; } -void yyset_lval (YYSTYPE * yylval_param , yyscan_t yyscanner) +void yyset_lval(YYSTYPE *yylval_param, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - yylval = yylval_param; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + yylval = yylval_param; } -YYLTYPE *yyget_lloc (yyscan_t yyscanner) +YYLTYPE *yyget_lloc(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - return yylloc; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + return yylloc; } - -void yyset_lloc (YYLTYPE * yylloc_param , yyscan_t yyscanner) + +void yyset_lloc(YYLTYPE *yylloc_param, yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - yylloc = yylloc_param; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + yylloc = yylloc_param; } - + /* User-visible API */ /* yylex_init is special because it creates the scanner itself, so it is * the ONLY reentrant function that doesn't take the scanner as the last argument. * That's why we explicitly handle the declaration, instead of using our macros. */ - -int yylex_init(yyscan_t* ptr_yy_globals) - +int yylex_init(yyscan_t *ptr_yy_globals) { - if (ptr_yy_globals == NULL){ - errno = EINVAL; - return 1; - } + if (ptr_yy_globals == NULL) { + errno = EINVAL; + return 1; + } - *ptr_yy_globals = (yyscan_t) yyalloc ( sizeof( struct yyguts_t ), NULL ); + *ptr_yy_globals = (yyscan_t)yyalloc(sizeof(struct yyguts_t), NULL); - if (*ptr_yy_globals == NULL){ - errno = ENOMEM; - return 1; - } + if (*ptr_yy_globals == NULL) { + errno = ENOMEM; + return 1; + } - /* By setting to 0xAA, we expose bugs in yy_init_globals. Leave at 0x00 for releases. */ - memset(*ptr_yy_globals,0x00,sizeof(struct yyguts_t)); + /* By setting to 0xAA, we expose bugs in yy_init_globals. Leave at 0x00 for releases. */ + memset(*ptr_yy_globals, 0x00, sizeof(struct yyguts_t)); - return yy_init_globals ( *ptr_yy_globals ); + return yy_init_globals(*ptr_yy_globals); } /* yylex_init_extra has the same functionality as yylex_init, but follows the @@ -2311,96 +4888,94 @@ int yylex_init(yyscan_t* ptr_yy_globals) * The user defined value in the first argument will be available to yyalloc in * the yyextra field. */ +int yylex_init_extra(YY_EXTRA_TYPE yy_user_defined, yyscan_t *ptr_yy_globals) +{ + struct yyguts_t dummy_yyguts; -int yylex_init_extra(YY_EXTRA_TYPE yy_user_defined,yyscan_t* ptr_yy_globals ) + yyset_extra(yy_user_defined, &dummy_yyguts); -{ - struct yyguts_t dummy_yyguts; + if (ptr_yy_globals == NULL) { + errno = EINVAL; + return 1; + } - yyset_extra (yy_user_defined, &dummy_yyguts); + *ptr_yy_globals = (yyscan_t)yyalloc(sizeof(struct yyguts_t), &dummy_yyguts); - if (ptr_yy_globals == NULL){ - errno = EINVAL; - return 1; - } - - *ptr_yy_globals = (yyscan_t) yyalloc ( sizeof( struct yyguts_t ), &dummy_yyguts ); - - if (*ptr_yy_globals == NULL){ - errno = ENOMEM; - return 1; - } - - /* By setting to 0xAA, we expose bugs in - yy_init_globals. Leave at 0x00 for releases. */ - memset(*ptr_yy_globals,0x00,sizeof(struct yyguts_t)); - - yyset_extra (yy_user_defined, *ptr_yy_globals); - - return yy_init_globals ( *ptr_yy_globals ); -} + if (*ptr_yy_globals == NULL) { + errno = ENOMEM; + return 1; + } -static int yy_init_globals (yyscan_t yyscanner) -{ - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - /* Initialization is the same as for the non-reentrant scanner. - * This function is called from yylex_destroy(), so don't allocate here. - */ + /* By setting to 0xAA, we expose bugs in + yy_init_globals. Leave at 0x00 for releases. */ + memset(*ptr_yy_globals, 0x00, sizeof(struct yyguts_t)); - yyg->yy_buffer_stack = 0; - yyg->yy_buffer_stack_top = 0; - yyg->yy_buffer_stack_max = 0; - yyg->yy_c_buf_p = (char *) 0; - yyg->yy_init = 0; - yyg->yy_start = 0; + yyset_extra(yy_user_defined, *ptr_yy_globals); - yyg->yy_start_stack_ptr = 0; - yyg->yy_start_stack_depth = 0; - yyg->yy_start_stack = NULL; + return yy_init_globals(*ptr_yy_globals); +} + +static int yy_init_globals(yyscan_t yyscanner) +{ + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + /* Initialization is the same as for the non-reentrant scanner. + * This function is called from yylex_destroy(), so don't allocate here. + */ + + yyg->yy_buffer_stack = NULL; + yyg->yy_buffer_stack_top = 0; + yyg->yy_buffer_stack_max = 0; + yyg->yy_c_buf_p = NULL; + yyg->yy_init = 0; + yyg->yy_start = 0; + + yyg->yy_start_stack_ptr = 0; + yyg->yy_start_stack_depth = 0; + yyg->yy_start_stack = NULL; /* Defined in main.c */ #ifdef YY_STDINIT - yyin = stdin; - yyout = stdout; + yyin = stdin; + yyout = stdout; #else - yyin = (FILE *) 0; - yyout = (FILE *) 0; + yyin = NULL; + yyout = NULL; #endif - /* For future reference: Set errno on error, since we are called by - * yylex_init() - */ - return 0; + /* For future reference: Set errno on error, since we are called by + * yylex_init() + */ + return 0; } /* yylex_destroy is for both reentrant and non-reentrant scanners. */ -int yylex_destroy (yyscan_t yyscanner) +int yylex_destroy(yyscan_t yyscanner) { - struct yyguts_t * yyg = (struct yyguts_t*)yyscanner; - - /* Pop the buffer stack, destroying each element. */ - while(YY_CURRENT_BUFFER){ - yy_delete_buffer(YY_CURRENT_BUFFER ,yyscanner ); - YY_CURRENT_BUFFER_LVALUE = NULL; - yypop_buffer_state(yyscanner); - } - - /* Destroy the stack itself. */ - yyfree(yyg->yy_buffer_stack ,yyscanner); - yyg->yy_buffer_stack = NULL; - - /* Destroy the start condition stack. */ - yyfree(yyg->yy_start_stack ,yyscanner ); - yyg->yy_start_stack = NULL; - - /* Reset the globals. This is important in a non-reentrant scanner so the next time - * yylex() is called, initialization will occur. */ - yy_init_globals( yyscanner); - - /* Destroy the main struct (reentrant only). */ - yyfree ( yyscanner , yyscanner ); - yyscanner = NULL; - return 0; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + + /* Pop the buffer stack, destroying each element. */ + while (YY_CURRENT_BUFFER) { + yy_delete_buffer(YY_CURRENT_BUFFER, yyscanner); + YY_CURRENT_BUFFER_LVALUE = NULL; + yypop_buffer_state(yyscanner); + } + + /* Destroy the stack itself. */ + yyfree(yyg->yy_buffer_stack, yyscanner); + yyg->yy_buffer_stack = NULL; + + /* Destroy the start condition stack. */ + yyfree(yyg->yy_start_stack, yyscanner); + yyg->yy_start_stack = NULL; + + /* Reset the globals. This is important in a non-reentrant scanner so the next time + * yylex() is called, initialization will occur. */ + yy_init_globals(yyscanner); + + /* Destroy the main struct (reentrant only). */ + yyfree(yyscanner, yyscanner); + yyscanner = NULL; + return 0; } /* @@ -2408,54 +4983,59 @@ int yylex_destroy (yyscan_t yyscanner) */ #ifndef yytext_ptr -static void yy_flex_strncpy (char* s1, yyconst char * s2, int n , yyscan_t yyscanner) +static void yy_flex_strncpy(char *s1, const char *s2, int n, yyscan_t yyscanner) { - register int i; - for ( i = 0; i < n; ++i ) - s1[i] = s2[i]; + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + (void)yyg; + + int i; + for (i = 0; i < n; ++i) + s1[i] = s2[i]; } #endif #ifdef YY_NEED_STRLEN -static int yy_flex_strlen (yyconst char * s , yyscan_t yyscanner) +static int yy_flex_strlen(const char *s, yyscan_t yyscanner) { - register int n; - for ( n = 0; s[n]; ++n ) - ; + int n; + for (n = 0; s[n]; ++n) + ; - return n; + return n; } #endif -void *yyalloc (yy_size_t size , yyscan_t yyscanner) +void *yyalloc(yy_size_t size, yyscan_t yyscanner) { - return (void *) malloc( size ); + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + (void)yyg; + return malloc(size); } -void *yyrealloc (void * ptr, yy_size_t size , yyscan_t yyscanner) +void *yyrealloc(void *ptr, yy_size_t size, yyscan_t yyscanner) { - /* The cast to (char *) in the following accommodates both - * implementations that use char* generic pointers, and those - * that use void* generic pointers. It works with the latter - * because both ANSI C and C++ allow castless assignment from - * any pointer type to void*, and deal with argument conversions - * as though doing an assignment. - */ - return (void *) realloc( (char *) ptr, size ); + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + (void)yyg; + + /* The cast to (char *) in the following accommodates both + * implementations that use char* generic pointers, and those + * that use void* generic pointers. It works with the latter + * because both ANSI C and C++ allow castless assignment from + * any pointer type to void*, and deal with argument conversions + * as though doing an assignment. + */ + return realloc(ptr, size); } -void yyfree (void * ptr , yyscan_t yyscanner) +void yyfree(void *ptr, yyscan_t yyscanner) { - free( (char *) ptr ); /* see yyrealloc() for (char *) cast */ + struct yyguts_t *yyg = (struct yyguts_t *)yyscanner; + (void)yyg; + free((char *)ptr); /* see yyrealloc() for (char *) cast */ } #define YYTABLES_NAME "yytables" -#line 143 "lex_sql.l" - - - -void scan_string(const char *str, yyscan_t scanner) { - yy_switch_to_buffer(yy_scan_string(str,scanner),scanner); -} +#line 171 "lex_sql.l" +void scan_string(const char *str, yyscan_t scanner) { yy_switch_to_buffer(yy_scan_string(str, scanner), scanner); } diff --git a/src/observer/sql/parser/lex_sql.h b/src/observer/sql/parser/lex_sql.h index 41767cf3..55c2c001 100644 --- a/src/observer/sql/parser/lex_sql.h +++ b/src/observer/sql/parser/lex_sql.h @@ -2,8 +2,7 @@ #define yyHEADER_H 1 #define yyIN_HEADER 1 -#line 6 "lex_sql.h" -#line 2 "lex_sql.l" +#line 5 "lex_sql.h" /* 这里的代码会被复制到lex_sql.cpp的最开始位置 定义yy_size_t的原因是因为flex生成的代码，会使用yy_size_t与其他类型的数字 @@ -13,37 +12,56 @@ typedef int yy_size_t; /* 参考生成的lex_sql.cpp代码，这个宏定义会放在每次运行yylex()最开始的地方 */ -#define YY_USER_INIT \ - yycolumn = 0; +#define YY_USER_INIT yycolumn = 0; /* 参考生成的lex_sql.cpp代码，这个宏定义会放在解析一个token之后，也可以在网上找到大量的参考资料 */ /* 我们在这里设置当前解析的token的位置信息，这样在yacc中就可以使用这些信息了 */ -#define YY_USER_ACTION \ -do { \ - yylloc->first_line = yylloc->last_line = yylineno; \ - yylloc->first_column = yycolumn; \ - yylloc->last_column = yylloc->first_column + yyleng - 1; \ - yycolumn += yyleng; \ -} \ -while (0); +#define YY_USER_ACTION \ + do { \ + yylloc->first_line = yylloc->last_line = yylineno; \ + yylloc->first_column = yycolumn; \ + yylloc->last_column = yylloc->first_column + yyleng - 1; \ + yycolumn += yyleng; \ + } while (0); +#line 29 "lex_sql.h" - - -#line 34 "lex_sql.h" - -#define YY_INT_ALIGNED short int +#define YY_INT_ALIGNED short int /* A lexical scanner generated by flex */ #define FLEX_SCANNER #define YY_FLEX_MAJOR_VERSION 2 -#define YY_FLEX_MINOR_VERSION 5 -#define YY_FLEX_SUBMINOR_VERSION 35 +#define YY_FLEX_MINOR_VERSION 6 +#define YY_FLEX_SUBMINOR_VERSION 4 #if YY_FLEX_SUBMINOR_VERSION > 0 #define FLEX_BETA #endif +#ifdef yyget_lval +#define yyget_lval_ALREADY_DEFINED +#else +#define yyget_lval yyget_lval +#endif + +#ifdef yyset_lval +#define yyset_lval_ALREADY_DEFINED +#else +#define yyset_lval yyset_lval +#endif + +#ifdef yyget_lloc +#define yyget_lloc_ALREADY_DEFINED +#else +#define yyget_lloc yyget_lloc +#endif + +#ifdef yyset_lloc +#define yyset_lloc_ALREADY_DEFINED +#else +#define yyset_lloc yyset_lloc +#endif + /* First, we deal with platform-specific or compiler-specific issues. */ /* begin standard C headers. */ @@ -61,87 +79,83 @@ while (0); /* C99 systems have . Non-C99 systems may or may not. */ -#if defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L +#if defined(__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* C99 says to define __STDC_LIMIT_MACROS before including stdint.h, - * if you want the limit (max/min) macros for int types. + * if you want the limit (max/min) macros for int types. */ #ifndef __STDC_LIMIT_MACROS #define __STDC_LIMIT_MACROS 1 #endif #include -typedef int8_t flex_int8_t; -typedef uint8_t flex_uint8_t; -typedef int16_t flex_int16_t; +typedef int8_t flex_int8_t; +typedef uint8_t flex_uint8_t; +typedef int16_t flex_int16_t; typedef uint16_t flex_uint16_t; -typedef int32_t flex_int32_t; +typedef int32_t flex_int32_t; typedef uint32_t flex_uint32_t; +typedef uint64_t flex_uint64_t; #else -typedef signed char flex_int8_t; -typedef short int flex_int16_t; -typedef int flex_int32_t; -typedef unsigned char flex_uint8_t; +typedef signed char flex_int8_t; +typedef short int flex_int16_t; +typedef int flex_int32_t; +typedef unsigned char flex_uint8_t; typedef unsigned short int flex_uint16_t; -typedef unsigned int flex_uint32_t; -#endif /* ! C99 */ +typedef unsigned int flex_uint32_t; /* Limits of integral types. */ #ifndef INT8_MIN -#define INT8_MIN (-128) +#define INT8_MIN (-128) #endif #ifndef INT16_MIN -#define INT16_MIN (-32767-1) +#define INT16_MIN (-32767 - 1) #endif #ifndef INT32_MIN -#define INT32_MIN (-2147483647-1) +#define INT32_MIN (-2147483647 - 1) #endif #ifndef INT8_MAX -#define INT8_MAX (127) +#define INT8_MAX (127) #endif #ifndef INT16_MAX -#define INT16_MAX (32767) +#define INT16_MAX (32767) #endif #ifndef INT32_MAX -#define INT32_MAX (2147483647) +#define INT32_MAX (2147483647) #endif #ifndef UINT8_MAX -#define UINT8_MAX (255U) +#define UINT8_MAX (255U) #endif #ifndef UINT16_MAX -#define UINT16_MAX (65535U) +#define UINT16_MAX (65535U) #endif #ifndef UINT32_MAX -#define UINT32_MAX (4294967295U) +#define UINT32_MAX (4294967295U) #endif -#endif /* ! FLEXINT_H */ - -#ifdef __cplusplus - -/* The "const" storage-class-modifier is valid. */ -#define YY_USE_CONST - -#else /* ! __cplusplus */ +#ifndef SIZE_MAX +#define SIZE_MAX (~(size_t)0) +#endif -/* C99 requires __STDC__ to be defined as 1. */ -#if defined (__STDC__) +#endif /* ! C99 */ -#define YY_USE_CONST +#endif /* ! FLEXINT_H */ -#endif /* defined (__STDC__) */ -#endif /* ! __cplusplus */ +/* begin standard C++ headers. */ -#ifdef YY_USE_CONST +/* TODO: this is always defined, so inline it */ #define yyconst const + +#if defined(__GNUC__) && __GNUC__ >= 3 +#define yynoreturn __attribute__((__noreturn__)) #else -#define yyconst +#define yynoreturn #endif /* An opaque pointer. */ #ifndef YY_TYPEDEF_YY_SCANNER_T #define YY_TYPEDEF_YY_SCANNER_T -typedef void* yyscan_t; +typedef void *yyscan_t; #endif /* For convenience, these vars (plus the bison vars far below) @@ -157,7 +171,15 @@ typedef void* yyscan_t; /* Size of default input buffer. */ #ifndef YY_BUF_SIZE +#ifdef __ia64__ +/* On IA-64, the buffer size is 16k, not 8k. + * Moreover, YY_BUF_SIZE is 2*YY_READ_BUF_SIZE in the general case. + * Ditto for the __ia64__ case accordingly. + */ +#define YY_BUF_SIZE 32768 +#else #define YY_BUF_SIZE 16384 +#endif /* __ia64__ */ #endif #ifndef YY_TYPEDEF_YY_BUFFER_STATE @@ -173,73 +195,72 @@ typedef size_t yy_size_t; #ifndef YY_STRUCT_YY_BUFFER_STATE #define YY_STRUCT_YY_BUFFER_STATE struct yy_buffer_state - { - FILE *yy_input_file; - - char *yy_ch_buf; /* input buffer */ - char *yy_buf_pos; /* current position in input buffer */ - - /* Size of input buffer in bytes, not including room for EOB - * characters. - */ - yy_size_t yy_buf_size; - - /* Number of characters read into yy_ch_buf, not including EOB - * characters. - */ - int yy_n_chars; - - /* Whether we "own" the buffer - i.e., we know we created it, - * and can realloc() it to grow it, and should free() it to - * delete it. - */ - int yy_is_our_buffer; - - /* Whether this is an "interactive" input source; if so, and - * if we're using stdio for input, then we want to use getc() - * instead of fread(), to make sure we stop fetching input after - * each newline. - */ - int yy_is_interactive; - - /* Whether we're considered to be at the beginning of a line. - * If so, '^' rules will be active on the next match, otherwise - * not. - */ - int yy_at_bol; - - int yy_bs_lineno; /**< The line count. */ - int yy_bs_column; /**< The column count. */ - - /* Whether to try to fill the input buffer when we reach the - * end of it. - */ - int yy_fill_buffer; - - int yy_buffer_status; - - }; +{ + FILE *yy_input_file; + + char *yy_ch_buf; /* input buffer */ + char *yy_buf_pos; /* current position in input buffer */ + + /* Size of input buffer in bytes, not including room for EOB + * characters. + */ + int yy_buf_size; + + /* Number of characters read into yy_ch_buf, not including EOB + * characters. + */ + yy_size_t yy_n_chars; + + /* Whether we "own" the buffer - i.e., we know we created it, + * and can realloc() it to grow it, and should free() it to + * delete it. + */ + int yy_is_our_buffer; + + /* Whether this is an "interactive" input source; if so, and + * if we're using stdio for input, then we want to use getc() + * instead of fread(), to make sure we stop fetching input after + * each newline. + */ + int yy_is_interactive; + + /* Whether we're considered to be at the beginning of a line. + * If so, '^' rules will be active on the next match, otherwise + * not. + */ + int yy_at_bol; + + int yy_bs_lineno; /**< The line count. */ + int yy_bs_column; /**< The column count. */ + + /* Whether to try to fill the input buffer when we reach the + * end of it. + */ + int yy_fill_buffer; + + int yy_buffer_status; +}; #endif /* !YY_STRUCT_YY_BUFFER_STATE */ -void yyrestart (FILE *input_file ,yyscan_t yyscanner ); -void yy_switch_to_buffer (YY_BUFFER_STATE new_buffer ,yyscan_t yyscanner ); -YY_BUFFER_STATE yy_create_buffer (FILE *file,int size ,yyscan_t yyscanner ); -void yy_delete_buffer (YY_BUFFER_STATE b ,yyscan_t yyscanner ); -void yy_flush_buffer (YY_BUFFER_STATE b ,yyscan_t yyscanner ); -void yypush_buffer_state (YY_BUFFER_STATE new_buffer ,yyscan_t yyscanner ); -void yypop_buffer_state (yyscan_t yyscanner ); +void yyrestart(FILE *input_file, yyscan_t yyscanner); +void yy_switch_to_buffer(YY_BUFFER_STATE new_buffer, yyscan_t yyscanner); +YY_BUFFER_STATE yy_create_buffer(FILE *file, int size, yyscan_t yyscanner); +void yy_delete_buffer(YY_BUFFER_STATE b, yyscan_t yyscanner); +void yy_flush_buffer(YY_BUFFER_STATE b, yyscan_t yyscanner); +void yypush_buffer_state(YY_BUFFER_STATE new_buffer, yyscan_t yyscanner); +void yypop_buffer_state(yyscan_t yyscanner); -YY_BUFFER_STATE yy_scan_buffer (char *base,yy_size_t size ,yyscan_t yyscanner ); -YY_BUFFER_STATE yy_scan_string (yyconst char *yy_str ,yyscan_t yyscanner ); -YY_BUFFER_STATE yy_scan_bytes (yyconst char *bytes,int len ,yyscan_t yyscanner ); +YY_BUFFER_STATE yy_scan_buffer(char *base, yy_size_t size, yyscan_t yyscanner); +YY_BUFFER_STATE yy_scan_string(const char *yy_str, yyscan_t yyscanner); +YY_BUFFER_STATE yy_scan_bytes(const char *bytes, yy_size_t len, yyscan_t yyscanner); -void *yyalloc (yy_size_t ,yyscan_t yyscanner ); -void *yyrealloc (void *,yy_size_t ,yyscan_t yyscanner ); -void yyfree (void * ,yyscan_t yyscanner ); +void *yyalloc(yy_size_t, yyscan_t yyscanner); +void *yyrealloc(void *, yy_size_t, yyscan_t yyscanner); +void yyfree(void *, yyscan_t yyscanner); /* Begin user sect3 */ -#define yywrap(n) 1 +#define yywrap(yyscanner) (/*CONSTCOND*/ 1) #define YY_SKIP_YYWRAP #define yytext_ptr yytext_r @@ -262,65 +283,69 @@ void yyfree (void * ,yyscan_t yyscanner ); #define YY_EXTRA_TYPE void * #endif -int yylex_init (yyscan_t* scanner); +int yylex_init(yyscan_t *scanner); -int yylex_init_extra (YY_EXTRA_TYPE user_defined,yyscan_t* scanner); +int yylex_init_extra(YY_EXTRA_TYPE user_defined, yyscan_t *scanner); /* Accessor methods to globals. These are made visible to non-reentrant scanners for convenience. */ -int yylex_destroy (yyscan_t yyscanner ); +int yylex_destroy(yyscan_t yyscanner); + +int yyget_debug(yyscan_t yyscanner); + +void yyset_debug(int debug_flag, yyscan_t yyscanner); + +YY_EXTRA_TYPE yyget_extra(yyscan_t yyscanner); -int yyget_debug (yyscan_t yyscanner ); +void yyset_extra(YY_EXTRA_TYPE user_defined, yyscan_t yyscanner); -void yyset_debug (int debug_flag ,yyscan_t yyscanner ); +FILE *yyget_in(yyscan_t yyscanner); -YY_EXTRA_TYPE yyget_extra (yyscan_t yyscanner ); +void yyset_in(FILE *_in_str, yyscan_t yyscanner); -void yyset_extra (YY_EXTRA_TYPE user_defined ,yyscan_t yyscanner ); +FILE *yyget_out(yyscan_t yyscanner); -FILE *yyget_in (yyscan_t yyscanner ); +void yyset_out(FILE *_out_str, yyscan_t yyscanner); -void yyset_in (FILE * in_str ,yyscan_t yyscanner ); +yy_size_t yyget_leng(yyscan_t yyscanner); -FILE *yyget_out (yyscan_t yyscanner ); +char *yyget_text(yyscan_t yyscanner); -void yyset_out (FILE * out_str ,yyscan_t yyscanner ); +int yyget_lineno(yyscan_t yyscanner); -int yyget_leng (yyscan_t yyscanner ); +void yyset_lineno(int _line_number, yyscan_t yyscanner); -char *yyget_text (yyscan_t yyscanner ); +int yyget_column(yyscan_t yyscanner); -int yyget_lineno (yyscan_t yyscanner ); +void yyset_column(int _column_no, yyscan_t yyscanner); -void yyset_lineno (int line_number ,yyscan_t yyscanner ); +YYSTYPE *yyget_lval(yyscan_t yyscanner); -YYSTYPE * yyget_lval (yyscan_t yyscanner ); +void yyset_lval(YYSTYPE *yylval_param, yyscan_t yyscanner); -void yyset_lval (YYSTYPE * yylval_param ,yyscan_t yyscanner ); +YYLTYPE *yyget_lloc(yyscan_t yyscanner); + +void yyset_lloc(YYLTYPE *yylloc_param, yyscan_t yyscanner); - YYLTYPE *yyget_lloc (yyscan_t yyscanner ); - - void yyset_lloc (YYLTYPE * yylloc_param ,yyscan_t yyscanner ); - /* Macros after this point can all be overridden by user definitions in * section 1. */ #ifndef YY_SKIP_YYWRAP #ifdef __cplusplus -extern "C" int yywrap (yyscan_t yyscanner ); +extern "C" int yywrap(yyscan_t yyscanner); #else -extern int yywrap (yyscan_t yyscanner ); +extern int yywrap(yyscan_t yyscanner); #endif #endif #ifndef yytext_ptr -static void yy_flex_strncpy (char *,yyconst char *,int ,yyscan_t yyscanner); +static void yy_flex_strncpy(char *, const char *, int, yyscan_t yyscanner); #endif #ifdef YY_NEED_STRLEN -static int yy_flex_strlen (yyconst char * ,yyscan_t yyscanner); +static int yy_flex_strlen(const char *, yyscan_t yyscanner); #endif #ifndef YY_NO_INPUT @@ -329,7 +354,12 @@ static int yy_flex_strlen (yyconst char * ,yyscan_t yyscanner); /* Amount of stuff to slurp up with each read. */ #ifndef YY_READ_BUF_SIZE +#ifdef __ia64__ +/* On IA-64, the buffer size is 16k, not 8k */ +#define YY_READ_BUF_SIZE 16384 +#else #define YY_READ_BUF_SIZE 8192 +#endif /* __ia64__ */ #endif /* Number of entries by which start-condition stack grows. */ @@ -343,11 +373,9 @@ static int yy_flex_strlen (yyconst char * ,yyscan_t yyscanner); #ifndef YY_DECL #define YY_DECL_IS_OURS 1 -extern int yylex \ - (YYSTYPE * yylval_param,YYLTYPE * yylloc_param ,yyscan_t yyscanner); +extern int yylex(YYSTYPE *yylval_param, YYLTYPE *yylloc_param, yyscan_t yyscanner); -#define YY_DECL int yylex \ - (YYSTYPE * yylval_param, YYLTYPE * yylloc_param , yyscan_t yyscanner) +#define YY_DECL int yylex(YYSTYPE *yylval_param, YYLTYPE *yylloc_param, yyscan_t yyscanner) #endif /* !YY_DECL */ /* yy_get_previous_state - get the state just before the EOB char was reached */ @@ -364,9 +392,153 @@ extern int yylex \ #undef YY_DECL #endif -#line 143 "lex_sql.l" +#ifndef yy_create_buffer_ALREADY_DEFINED +#undef yy_create_buffer +#endif +#ifndef yy_delete_buffer_ALREADY_DEFINED +#undef yy_delete_buffer +#endif +#ifndef yy_scan_buffer_ALREADY_DEFINED +#undef yy_scan_buffer +#endif +#ifndef yy_scan_string_ALREADY_DEFINED +#undef yy_scan_string +#endif +#ifndef yy_scan_bytes_ALREADY_DEFINED +#undef yy_scan_bytes +#endif +#ifndef yy_init_buffer_ALREADY_DEFINED +#undef yy_init_buffer +#endif +#ifndef yy_flush_buffer_ALREADY_DEFINED +#undef yy_flush_buffer +#endif +#ifndef yy_load_buffer_state_ALREADY_DEFINED +#undef yy_load_buffer_state +#endif +#ifndef yy_switch_to_buffer_ALREADY_DEFINED +#undef yy_switch_to_buffer +#endif +#ifndef yypush_buffer_state_ALREADY_DEFINED +#undef yypush_buffer_state +#endif +#ifndef yypop_buffer_state_ALREADY_DEFINED +#undef yypop_buffer_state +#endif +#ifndef yyensure_buffer_stack_ALREADY_DEFINED +#undef yyensure_buffer_stack +#endif +#ifndef yylex_ALREADY_DEFINED +#undef yylex +#endif +#ifndef yyrestart_ALREADY_DEFINED +#undef yyrestart +#endif +#ifndef yylex_init_ALREADY_DEFINED +#undef yylex_init +#endif +#ifndef yylex_init_extra_ALREADY_DEFINED +#undef yylex_init_extra +#endif +#ifndef yylex_destroy_ALREADY_DEFINED +#undef yylex_destroy +#endif +#ifndef yyget_debug_ALREADY_DEFINED +#undef yyget_debug +#endif +#ifndef yyset_debug_ALREADY_DEFINED +#undef yyset_debug +#endif +#ifndef yyget_extra_ALREADY_DEFINED +#undef yyget_extra +#endif +#ifndef yyset_extra_ALREADY_DEFINED +#undef yyset_extra +#endif +#ifndef yyget_in_ALREADY_DEFINED +#undef yyget_in +#endif +#ifndef yyset_in_ALREADY_DEFINED +#undef yyset_in +#endif +#ifndef yyget_out_ALREADY_DEFINED +#undef yyget_out +#endif +#ifndef yyset_out_ALREADY_DEFINED +#undef yyset_out +#endif +#ifndef yyget_leng_ALREADY_DEFINED +#undef yyget_leng +#endif +#ifndef yyget_text_ALREADY_DEFINED +#undef yyget_text +#endif +#ifndef yyget_lineno_ALREADY_DEFINED +#undef yyget_lineno +#endif +#ifndef yyset_lineno_ALREADY_DEFINED +#undef yyset_lineno +#endif +#ifndef yyget_column_ALREADY_DEFINED +#undef yyget_column +#endif +#ifndef yyset_column_ALREADY_DEFINED +#undef yyset_column +#endif +#ifndef yywrap_ALREADY_DEFINED +#undef yywrap +#endif +#ifndef yyget_lval_ALREADY_DEFINED +#undef yyget_lval +#endif +#ifndef yyset_lval_ALREADY_DEFINED +#undef yyset_lval +#endif +#ifndef yyget_lloc_ALREADY_DEFINED +#undef yyget_lloc +#endif +#ifndef yyset_lloc_ALREADY_DEFINED +#undef yyset_lloc +#endif +#ifndef yyalloc_ALREADY_DEFINED +#undef yyalloc +#endif +#ifndef yyrealloc_ALREADY_DEFINED +#undef yyrealloc +#endif +#ifndef yyfree_ALREADY_DEFINED +#undef yyfree +#endif +#ifndef yytext_ALREADY_DEFINED +#undef yytext +#endif +#ifndef yyleng_ALREADY_DEFINED +#undef yyleng +#endif +#ifndef yyin_ALREADY_DEFINED +#undef yyin +#endif +#ifndef yyout_ALREADY_DEFINED +#undef yyout +#endif +#ifndef yy_flex_debug_ALREADY_DEFINED +#undef yy_flex_debug +#endif +#ifndef yylineno_ALREADY_DEFINED +#undef yylineno +#endif +#ifndef yytables_fload_ALREADY_DEFINED +#undef yytables_fload +#endif +#ifndef yytables_destroy_ALREADY_DEFINED +#undef yytables_destroy +#endif +#ifndef yyTABLES_NAME_ALREADY_DEFINED +#undef yyTABLES_NAME +#endif +#line 171 "lex_sql.l" -#line 371 "lex_sql.h" +#line 548 "lex_sql.h" #undef yyIN_HEADER #endif /* yyHEADER_H */ diff --git a/src/observer/sql/parser/lex_sql.l b/src/observer/sql/parser/lex_sql.l index 505c5ffe..721d3b66 100644 --- a/src/observer/sql/parser/lex_sql.l +++ b/src/observer/sql/parser/lex_sql.l @@ -77,8 +77,8 @@ QUOTE [\'\"] {WHITE_SAPCE} // ignore whitespace \n ; -[\-]?{DIGIT}+ yylval->number=atoi(yytext); RETURN_TOKEN(NUMBER); -[\-]?{DIGIT}+{DOT}{DIGIT}+ yylval->floats=(float)(atof(yytext)); RETURN_TOKEN(FLOAT); +{DIGIT}+ yylval->number=atoi(yytext); RETURN_TOKEN(NUMBER); +{DIGIT}+{DOT}{DIGIT}+ yylval->floats=(float)(atof(yytext)); RETURN_TOKEN(FLOAT); ";" RETURN_TOKEN(SEMICOLON); {DOT} RETURN_TOKEN(DOT); @@ -98,6 +98,7 @@ CALC RETURN_TOKEN(CALC); FROM RETURN_TOKEN(FROM); WHERE RETURN_TOKEN(WHERE); AND RETURN_TOKEN(AND); +OR RETURN_TOKEN(OR); INSERT RETURN_TOKEN(INSERT); INTO RETURN_TOKEN(INTO); VALUES RETURN_TOKEN(VALUES); @@ -110,18 +111,40 @@ ROLLBACK RETURN_TOKEN(TRX_ROLLBACK); INT RETURN_TOKEN(INT_T); CHAR RETURN_TOKEN(STRING_T); FLOAT RETURN_TOKEN(FLOAT_T); +DATE RETURN_TOKEN(DATE_T); +TEXT RETURN_TOKEN(TEXT_T); VECTOR RETURN_TOKEN(VECTOR_T); -LOAD RETURN_TOKEN(LOAD); -DATA RETURN_TOKEN(DATA); +UNIQUE RETURN_TOKEN(UNIQUE); +NULL RETURN_TOKEN(NULL_T); +NULLABLE RETURN_TOKEN(NULLABLE); INFILE RETURN_TOKEN(INFILE); EXPLAIN RETURN_TOKEN(EXPLAIN); GROUP RETURN_TOKEN(GROUP); +LIMIT RETURN_TOKEN(LIMIT); +HAVING RETURN_TOKEN(HAVING); +ORDER RETURN_TOKEN(ORDER); BY RETURN_TOKEN(BY); +AS RETURN_TOKEN(AS); +ASC RETURN_TOKEN(ASC); +IN RETURN_TOKEN(IN); +EXISTS RETURN_TOKEN(EXISTS); STORAGE RETURN_TOKEN(STORAGE); FORMAT RETURN_TOKEN(FORMAT); -{ID} yylval->string=strdup(yytext); RETURN_TOKEN(ID); +INNER RETURN_TOKEN(INNER); +JOIN RETURN_TOKEN(JOIN); +VIEW RETURN_TOKEN(VIEW); +WITH RETURN_TOKEN(WITH); +DISTANCE RETURN_TOKEN(DISTANCE); +TYPE RETURN_TOKEN(TYPE); +LISTS RETURN_TOKEN(LISTS); +PROBES RETURN_TOKEN(PROBES); +IVFFLAT RETURN_TOKEN(IVFFLAT); +TRUE RETURN_TOKEN(TRUE); +FALSE RETURN_TOKEN(FALSE); "(" RETURN_TOKEN(LBRACE); ")" RETURN_TOKEN(RBRACE); +"[" RETURN_TOKEN(LSBRACE); +"]" RETURN_TOKEN(RSBRACE); "," RETURN_TOKEN(COMMA); "=" RETURN_TOKEN(EQ); @@ -131,6 +154,11 @@ FORMAT RETURN_TOKEN(FORMAT); "<" RETURN_TOKEN(LT); ">=" RETURN_TOKEN(GE); ">" RETURN_TOKEN(GT); +NOT RETURN_TOKEN(NOT); +IS RETURN_TOKEN(IS); +LIKE RETURN_TOKEN(LIKE); + +{ID} yylval->string=strdup(yytext); RETURN_TOKEN(ID); "+" | "-" | diff --git a/src/observer/sql/parser/parse_defs.h b/src/observer/sql/parser/parse_defs.h index 9593d57a..57e251f0 100644 --- a/src/observer/sql/parser/parse_defs.h +++ b/src/observer/sql/parser/parse_defs.h @@ -15,6 +15,7 @@ See the Mulan PSL v2 for more details. */ #pragma once #include +// #include #include #include @@ -45,12 +46,20 @@ struct RelAttrSqlNode */ enum CompOp { - EQUAL_TO, ///< "=" - LESS_EQUAL, ///< "<=" - NOT_EQUAL, ///< "<>" - LESS_THAN, ///< "<" - GREAT_EQUAL, ///< ">=" - GREAT_THAN, ///< ">" + EQUAL_TO, ///< "=" + LESS_EQUAL, ///< "<=" + NOT_EQUAL, ///< "<>" + LESS_THAN, ///< "<" + GREAT_EQUAL, ///< ">=" + GREAT_THAN, ///< ">" + IS_OP, ///< "is (null)" + IS_NOT_OP, ///< "is not (null)" + LIKE_OP, ///< "like" + NOT_LIKE_OP, ///< "not like" + IN_OP, ///< "in (sub query)" + NOT_IN_OP, ///< "not in (sub query)" + EXISTS_OP, ///< "exists (sub query)" + NOT_EXISTS_OP, ///< "not exists (sub query)" NO_OP }; @@ -75,6 +84,31 @@ struct ConditionSqlNode Value right_value; ///< right-hand side value if right_is_attr = FALSE }; +/** + * @brief 描述一个relation的节点 + */ +struct RelationNode +{ + RelationNode(std::string relation_, std::string alias_) : relation(std::move(relation_)), alias(std::move(alias_)) {} + explicit RelationNode(std::string relation_) : relation(std::move(relation_)) {} + std::string relation; ///< 查询的表 + std::string alias; ///< 该表的别名 (may be NULL) +}; + +/** + * @brief 描述一个orderby的节点 + */ +struct OrderBySqlNode +{ + std::unique_ptr expr; + bool is_asc; ///< 默认true 为升序 +}; + +struct LimitSqlNode +{ + int number; +}; + /** * @brief 描述一个select语句 * @ingroup SQLParser @@ -85,13 +119,26 @@ struct ConditionSqlNode * where 条件 conditions，这里表示使用AND串联起来多个条件。正常的SQL语句会有OR，NOT等， * 甚至可以包含复杂的表达式。 */ - struct SelectSqlNode { - std::vector> expressions; ///< 查询的表达式 - std::vector relations; ///< 查询的表 - std::vector conditions; ///< 查询条件，使用AND串联起来多个条件 - std::vector> group_by; ///< group by clause + std::vector> expressions; ///< 查询的表达式 + std::vector relations; ///< 查询的表 + std::unique_ptr conditions; ///< 查询条件，使用AND串联起来多个条件 + std::vector> group_by; ///< group by clause + std::vector order_by; ///< attributes in order clause + std::unique_ptr having_conditions; ///< having + std::unique_ptr limit; +}; + +/** + * @brief 描述一个join语句 + * @ingroup SQLParser + * @details 目前只支持 inner join，解析表和条件后直接放到 SelectSqlNode 里。 + */ +struct JoinSqlNode +{ + std::vector relations; ///< 查询的表 + std::unique_ptr conditions; ///< 查询条件，可能有多个 }; /** @@ -110,8 +157,9 @@ struct CalcSqlNode */ struct InsertSqlNode { - std::string relation_name; ///< Relation to insert into - std::vector values; ///< 要插入的值 + std::string relation_name; ///< Relation to insert into + std::vector attr_names; ///< attribute names + std::vector> values_list; ///< 要插入的值列表 }; /** @@ -120,8 +168,18 @@ struct InsertSqlNode */ struct DeleteSqlNode { - std::string relation_name; ///< Relation to delete from - std::vector conditions; + std::string relation_name; ///< Relation to delete from + std::unique_ptr condition; +}; + +/** + * @brief 描述一个set语句 + * @ingroup SQLParser + */ +struct SetClauseSqlNode +{ + std::string field_name; ///< 更新的字段 + std::unique_ptr value; ///< 更新的值 }; /** @@ -130,10 +188,9 @@ struct DeleteSqlNode */ struct UpdateSqlNode { - std::string relation_name; ///< Relation to update - std::string attribute_name; ///< 更新的字段，仅支持一个字段 - Value value; ///< 更新的值，仅支持一个字段 - std::vector conditions; + std::string relation_name; ///< Relation to update + std::vector set_clauses; ///< 更新的set语句，支持多个字段和值 + std::unique_ptr conditions; ///< 谓词条件 }; /** @@ -143,9 +200,11 @@ struct UpdateSqlNode */ struct AttrInfoSqlNode { - AttrType type; ///< Type of attribute - std::string name; ///< Attribute name - size_t length; ///< Length of attribute + AttrType type; ///< Type of attribute + std::string name; ///< Attribute name + size_t length; ///< Length of attribute + bool nullable; ///< 字段是否可以为空 + bool mutable_ = true; ///< 视图字段是否是可插入修改的 }; /** @@ -155,9 +214,32 @@ struct AttrInfoSqlNode */ struct CreateTableSqlNode { - std::string relation_name; ///< Relation name - std::vector attr_infos; ///< attributes - std::string storage_format; ///< storage format + std::string relation_name; ///< Relation name + std::vector attr_infos; ///< attributes + std::string storage_format; ///< storage format + std::unique_ptr create_table_select; ///< create table select +}; + +/** + * @brief 描述一个create view语句 + * @ingroup SQLParser + * @details 这里也做了很多简化。 + */ +struct CreateViewSqlNode +{ + std::string relation_name; ///< Relation name + std::vector attribute_names; ///< attribute names + std::unique_ptr create_view_select; ///< create table select +}; + +/** + * @brief 描述一个drop view语句 + * @ingroup SQLParser + * @details 这里也做了很多简化。 + */ +struct DropViewSqlNode +{ + std::string relation_name; ///< Relation name }; /** @@ -169,6 +251,21 @@ struct DropTableSqlNode std::string relation_name; ///< 要删除的表名 }; +enum class IndexType +{ + BPlusTreeIndex, + VectorIVFFlatIndex, // 目前就支持以上两种 + VectorHNSWIndex +}; + +struct VectorIndexConfig +{ + std::string distance_fn; ///< 距离度量方法 + IndexType index_type = IndexType::BPlusTreeIndex; ///< 索引类型，非向量索引情况默认为 b+ 树 + Value lists; ///< 列表数量 + Value probes; ///< 探测次数 +}; + /** * @brief 描述一个create index语句 * @ingroup SQLParser @@ -177,9 +274,11 @@ struct DropTableSqlNode */ struct CreateIndexSqlNode { - std::string index_name; ///< Index name - std::string relation_name; ///< Relation name - std::string attribute_name; ///< Attribute name + bool unique; ///< unique index + std::string index_name; ///< Index name + std::string relation_name; ///< Relation name + std::vector attribute_name; ///< Attribute name + VectorIndexConfig vector_index_config; ///< 向量索引还有一些额外参数 }; /** @@ -192,6 +291,15 @@ struct DropIndexSqlNode std::string relation_name; ///< Relation name }; +/** + * @brief 描述一个show index语句 + * @ingroup SQLParser + */ +struct ShowIndexSqlNode +{ + std::string relation_name; ///< Relation name +}; + /** * @brief 描述一个desc table语句 * @ingroup SQLParser @@ -266,9 +374,12 @@ enum SqlCommandFlag SCF_DROP_TABLE, SCF_CREATE_INDEX, SCF_DROP_INDEX, + SCF_SHOW_INDEX, SCF_SYNC, SCF_SHOW_TABLES, SCF_DESC_TABLE, + SCF_CREATE_VIEW, + SCF_DROP_VIEW, SCF_BEGIN, ///< 事务开始语句，可以在这里扩展只读事务 SCF_COMMIT, SCF_CLOG_SYNC, @@ -297,7 +408,10 @@ class ParsedSqlNode DropTableSqlNode drop_table; CreateIndexSqlNode create_index; DropIndexSqlNode drop_index; + ShowIndexSqlNode show_index; DescTableSqlNode desc_table; + CreateViewSqlNode create_view; + DropViewSqlNode drop_view; LoadDataSqlNode load_data; ExplainSqlNode explain; SetVariableSqlNode set_variable; diff --git a/src/observer/sql/parser/yacc_sql.cpp b/src/observer/sql/parser/yacc_sql.cpp index becf04d6..ad0178e6 100644 --- a/src/observer/sql/parser/yacc_sql.cpp +++ b/src/observer/sql/parser/yacc_sql.cpp @@ -63,13 +63,9 @@ /* Pull parsers. */ #define YYPULL 1 - - - /* First part of user prologue. */ #line 2 "yacc_sql.y" - #include #include #include @@ -92,172 +88,239 @@ string token_name(const char *sql_string, YYLTYPE *llocp) int yyerror(YYLTYPE *llocp, const char *sql_string, ParsedSqlResult *sql_result, yyscan_t scanner, const char *msg) { std::unique_ptr error_sql_node = std::make_unique(SCF_ERROR); - error_sql_node->error.error_msg = msg; - error_sql_node->error.line = llocp->first_line; - error_sql_node->error.column = llocp->first_column; + error_sql_node->error.error_msg = msg; + error_sql_node->error.line = llocp->first_line; + error_sql_node->error.column = llocp->first_column; sql_result->add_sql_node(std::move(error_sql_node)); return 0; } -ArithmeticExpr *create_arithmetic_expression(ArithmeticExpr::Type type, - Expression *left, - Expression *right, - const char *sql_string, - YYLTYPE *llocp) +ArithmeticExpr *create_arithmetic_expression( + ArithmeticExpr::Type type, Expression *left, Expression *right, const char *sql_string, YYLTYPE *llocp) { ArithmeticExpr *expr = new ArithmeticExpr(type, left, right); expr->set_name(token_name(sql_string, llocp)); return expr; } -UnboundAggregateExpr *create_aggregate_expression(const char *aggregate_name, - Expression *child, - const char *sql_string, - YYLTYPE *llocp) +UnboundFunctionExpr *create_aggregate_expression( + const char *function_name, std::vector> child, const char *sql_string, YYLTYPE *llocp) { - UnboundAggregateExpr *expr = new UnboundAggregateExpr(aggregate_name, child); + UnboundFunctionExpr *expr = new UnboundFunctionExpr(function_name, std::move(child)); expr->set_name(token_name(sql_string, llocp)); return expr; } +ParsedSqlNode *create_table_sql_node(char *table_name, AttrInfoSqlNode *attr_def, + std::vector *attrinfos, char *storage_format, ParsedSqlNode *create_table_select) +{ + ParsedSqlNode *parsed_sql_node = new ParsedSqlNode(SCF_CREATE_TABLE); + CreateTableSqlNode &create_table = parsed_sql_node->create_table; + create_table.relation_name = table_name; + + if (attrinfos) { + create_table.attr_infos.swap(*attrinfos); + delete attrinfos; + } + if (attr_def) { + create_table.attr_infos.emplace_back(*attr_def); + std::reverse(create_table.attr_infos.begin(), create_table.attr_infos.end()); + delete attr_def; + } + if (storage_format != nullptr) { + create_table.storage_format = storage_format; + free(storage_format); + } + + if (create_table_select) { + create_table.create_table_select = std::make_unique(std::move(create_table_select->selection)); + } + + return parsed_sql_node; +} + +#line 155 "yacc_sql.cpp" -#line 125 "yacc_sql.cpp" - -# ifndef YY_CAST -# ifdef __cplusplus -# define YY_CAST(Type, Val) static_cast (Val) -# define YY_REINTERPRET_CAST(Type, Val) reinterpret_cast (Val) -# else -# define YY_CAST(Type, Val) ((Type) (Val)) -# define YY_REINTERPRET_CAST(Type, Val) ((Type) (Val)) -# endif -# endif -# ifndef YY_NULLPTR -# if defined __cplusplus -# if 201103L <= __cplusplus -# define YY_NULLPTR nullptr -# else -# define YY_NULLPTR 0 -# endif -# else -# define YY_NULLPTR ((void*)0) -# endif -# endif +#ifndef YY_CAST +#ifdef __cplusplus +#define YY_CAST(Type, Val) static_cast(Val) +#define YY_REINTERPRET_CAST(Type, Val) reinterpret_cast(Val) +#else +#define YY_CAST(Type, Val) ((Type)(Val)) +#define YY_REINTERPRET_CAST(Type, Val) ((Type)(Val)) +#endif +#endif +#ifndef YY_NULLPTR +#if defined __cplusplus +#if 201103L <= __cplusplus +#define YY_NULLPTR nullptr +#else +#define YY_NULLPTR 0 +#endif +#else +#define YY_NULLPTR ((void *)0) +#endif +#endif #include "yacc_sql.hpp" /* Symbol kind. */ enum yysymbol_kind_t { - YYSYMBOL_YYEMPTY = -2, - YYSYMBOL_YYEOF = 0, /* "end of file" */ - YYSYMBOL_YYerror = 1, /* error */ - YYSYMBOL_YYUNDEF = 2, /* "invalid token" */ - YYSYMBOL_SEMICOLON = 3, /* SEMICOLON */ - YYSYMBOL_BY = 4, /* BY */ - YYSYMBOL_CREATE = 5, /* CREATE */ - YYSYMBOL_DROP = 6, /* DROP */ - YYSYMBOL_GROUP = 7, /* GROUP */ - YYSYMBOL_TABLE = 8, /* TABLE */ - YYSYMBOL_TABLES = 9, /* TABLES */ - YYSYMBOL_INDEX = 10, /* INDEX */ - YYSYMBOL_CALC = 11, /* CALC */ - YYSYMBOL_SELECT = 12, /* SELECT */ - YYSYMBOL_DESC = 13, /* DESC */ - YYSYMBOL_SHOW = 14, /* SHOW */ - YYSYMBOL_SYNC = 15, /* SYNC */ - YYSYMBOL_INSERT = 16, /* INSERT */ - YYSYMBOL_DELETE = 17, /* DELETE */ - YYSYMBOL_UPDATE = 18, /* UPDATE */ - YYSYMBOL_LBRACE = 19, /* LBRACE */ - YYSYMBOL_RBRACE = 20, /* RBRACE */ - YYSYMBOL_COMMA = 21, /* COMMA */ - YYSYMBOL_TRX_BEGIN = 22, /* TRX_BEGIN */ - YYSYMBOL_TRX_COMMIT = 23, /* TRX_COMMIT */ - YYSYMBOL_TRX_ROLLBACK = 24, /* TRX_ROLLBACK */ - YYSYMBOL_INT_T = 25, /* INT_T */ - YYSYMBOL_STRING_T = 26, /* STRING_T */ - YYSYMBOL_FLOAT_T = 27, /* FLOAT_T */ - YYSYMBOL_VECTOR_T = 28, /* VECTOR_T */ - YYSYMBOL_HELP = 29, /* HELP */ - YYSYMBOL_EXIT = 30, /* EXIT */ - YYSYMBOL_DOT = 31, /* DOT */ - YYSYMBOL_INTO = 32, /* INTO */ - YYSYMBOL_VALUES = 33, /* VALUES */ - YYSYMBOL_FROM = 34, /* FROM */ - YYSYMBOL_WHERE = 35, /* WHERE */ - YYSYMBOL_AND = 36, /* AND */ - YYSYMBOL_SET = 37, /* SET */ - YYSYMBOL_ON = 38, /* ON */ - YYSYMBOL_LOAD = 39, /* LOAD */ - YYSYMBOL_DATA = 40, /* DATA */ - YYSYMBOL_INFILE = 41, /* INFILE */ - YYSYMBOL_EXPLAIN = 42, /* EXPLAIN */ - YYSYMBOL_STORAGE = 43, /* STORAGE */ - YYSYMBOL_FORMAT = 44, /* FORMAT */ - YYSYMBOL_EQ = 45, /* EQ */ - YYSYMBOL_LT = 46, /* LT */ - YYSYMBOL_GT = 47, /* GT */ - YYSYMBOL_LE = 48, /* LE */ - YYSYMBOL_GE = 49, /* GE */ - YYSYMBOL_NE = 50, /* NE */ - YYSYMBOL_NUMBER = 51, /* NUMBER */ - YYSYMBOL_FLOAT = 52, /* FLOAT */ - YYSYMBOL_ID = 53, /* ID */ - YYSYMBOL_SSS = 54, /* SSS */ - YYSYMBOL_55_ = 55, /* '+' */ - YYSYMBOL_56_ = 56, /* '-' */ - YYSYMBOL_57_ = 57, /* '*' */ - YYSYMBOL_58_ = 58, /* '/' */ - YYSYMBOL_UMINUS = 59, /* UMINUS */ - YYSYMBOL_YYACCEPT = 60, /* $accept */ - YYSYMBOL_commands = 61, /* commands */ - YYSYMBOL_command_wrapper = 62, /* command_wrapper */ - YYSYMBOL_exit_stmt = 63, /* exit_stmt */ - YYSYMBOL_help_stmt = 64, /* help_stmt */ - YYSYMBOL_sync_stmt = 65, /* sync_stmt */ - YYSYMBOL_begin_stmt = 66, /* begin_stmt */ - YYSYMBOL_commit_stmt = 67, /* commit_stmt */ - YYSYMBOL_rollback_stmt = 68, /* rollback_stmt */ - YYSYMBOL_drop_table_stmt = 69, /* drop_table_stmt */ - YYSYMBOL_show_tables_stmt = 70, /* show_tables_stmt */ - YYSYMBOL_desc_table_stmt = 71, /* desc_table_stmt */ - YYSYMBOL_create_index_stmt = 72, /* create_index_stmt */ - YYSYMBOL_drop_index_stmt = 73, /* drop_index_stmt */ - YYSYMBOL_create_table_stmt = 74, /* create_table_stmt */ - YYSYMBOL_attr_def_list = 75, /* attr_def_list */ - YYSYMBOL_attr_def = 76, /* attr_def */ - YYSYMBOL_number = 77, /* number */ - YYSYMBOL_type = 78, /* type */ - YYSYMBOL_insert_stmt = 79, /* insert_stmt */ - YYSYMBOL_value_list = 80, /* value_list */ - YYSYMBOL_value = 81, /* value */ - YYSYMBOL_storage_format = 82, /* storage_format */ - YYSYMBOL_delete_stmt = 83, /* delete_stmt */ - YYSYMBOL_update_stmt = 84, /* update_stmt */ - YYSYMBOL_select_stmt = 85, /* select_stmt */ - YYSYMBOL_calc_stmt = 86, /* calc_stmt */ - YYSYMBOL_expression_list = 87, /* expression_list */ - YYSYMBOL_expression = 88, /* expression */ - YYSYMBOL_rel_attr = 89, /* rel_attr */ - YYSYMBOL_relation = 90, /* relation */ - YYSYMBOL_rel_list = 91, /* rel_list */ - YYSYMBOL_where = 92, /* where */ - YYSYMBOL_condition_list = 93, /* condition_list */ - YYSYMBOL_condition = 94, /* condition */ - YYSYMBOL_comp_op = 95, /* comp_op */ - YYSYMBOL_group_by = 96, /* group_by */ - YYSYMBOL_load_data_stmt = 97, /* load_data_stmt */ - YYSYMBOL_explain_stmt = 98, /* explain_stmt */ - YYSYMBOL_set_variable_stmt = 99, /* set_variable_stmt */ - YYSYMBOL_opt_semicolon = 100 /* opt_semicolon */ + YYSYMBOL_YYEMPTY = -2, + YYSYMBOL_YYEOF = 0, /* "end of file" */ + YYSYMBOL_YYerror = 1, /* error */ + YYSYMBOL_YYUNDEF = 2, /* "invalid token" */ + YYSYMBOL_SEMICOLON = 3, /* SEMICOLON */ + YYSYMBOL_AS = 4, /* AS */ + YYSYMBOL_ASC = 5, /* ASC */ + YYSYMBOL_BY = 6, /* BY */ + YYSYMBOL_CREATE = 7, /* CREATE */ + YYSYMBOL_DROP = 8, /* DROP */ + YYSYMBOL_EXISTS = 9, /* EXISTS */ + YYSYMBOL_GROUP = 10, /* GROUP */ + YYSYMBOL_HAVING = 11, /* HAVING */ + YYSYMBOL_ORDER = 12, /* ORDER */ + YYSYMBOL_TABLE = 13, /* TABLE */ + YYSYMBOL_TABLES = 14, /* TABLES */ + YYSYMBOL_INDEX = 15, /* INDEX */ + YYSYMBOL_CALC = 16, /* CALC */ + YYSYMBOL_SELECT = 17, /* SELECT */ + YYSYMBOL_DESC = 18, /* DESC */ + YYSYMBOL_SHOW = 19, /* SHOW */ + YYSYMBOL_SYNC = 20, /* SYNC */ + YYSYMBOL_INSERT = 21, /* INSERT */ + YYSYMBOL_DELETE = 22, /* DELETE */ + YYSYMBOL_UPDATE = 23, /* UPDATE */ + YYSYMBOL_LBRACE = 24, /* LBRACE */ + YYSYMBOL_RBRACE = 25, /* RBRACE */ + YYSYMBOL_LSBRACE = 26, /* LSBRACE */ + YYSYMBOL_RSBRACE = 27, /* RSBRACE */ + YYSYMBOL_COMMA = 28, /* COMMA */ + YYSYMBOL_TRX_BEGIN = 29, /* TRX_BEGIN */ + YYSYMBOL_TRX_COMMIT = 30, /* TRX_COMMIT */ + YYSYMBOL_TRX_ROLLBACK = 31, /* TRX_ROLLBACK */ + YYSYMBOL_INT_T = 32, /* INT_T */ + YYSYMBOL_IN = 33, /* IN */ + YYSYMBOL_TRUE = 34, /* TRUE */ + YYSYMBOL_FALSE = 35, /* FALSE */ + YYSYMBOL_STRING_T = 36, /* STRING_T */ + YYSYMBOL_FLOAT_T = 37, /* FLOAT_T */ + YYSYMBOL_DATE_T = 38, /* DATE_T */ + YYSYMBOL_TEXT_T = 39, /* TEXT_T */ + YYSYMBOL_VECTOR_T = 40, /* VECTOR_T */ + YYSYMBOL_NOT = 41, /* NOT */ + YYSYMBOL_UNIQUE = 42, /* UNIQUE */ + YYSYMBOL_NULL_T = 43, /* NULL_T */ + YYSYMBOL_LIMIT = 44, /* LIMIT */ + YYSYMBOL_NULLABLE = 45, /* NULLABLE */ + YYSYMBOL_HELP = 46, /* HELP */ + YYSYMBOL_QUOTE = 47, /* QUOTE */ + YYSYMBOL_EXIT = 48, /* EXIT */ + YYSYMBOL_DOT = 49, /* DOT */ + YYSYMBOL_INTO = 50, /* INTO */ + YYSYMBOL_VALUES = 51, /* VALUES */ + YYSYMBOL_FROM = 52, /* FROM */ + YYSYMBOL_WHERE = 53, /* WHERE */ + YYSYMBOL_AND = 54, /* AND */ + YYSYMBOL_OR = 55, /* OR */ + YYSYMBOL_SET = 56, /* SET */ + YYSYMBOL_ON = 57, /* ON */ + YYSYMBOL_INFILE = 58, /* INFILE */ + YYSYMBOL_EXPLAIN = 59, /* EXPLAIN */ + YYSYMBOL_STORAGE = 60, /* STORAGE */ + YYSYMBOL_FORMAT = 61, /* FORMAT */ + YYSYMBOL_INNER = 62, /* INNER */ + YYSYMBOL_JOIN = 63, /* JOIN */ + YYSYMBOL_VIEW = 64, /* VIEW */ + YYSYMBOL_WITH = 65, /* WITH */ + YYSYMBOL_DISTANCE = 66, /* DISTANCE */ + YYSYMBOL_TYPE = 67, /* TYPE */ + YYSYMBOL_LISTS = 68, /* LISTS */ + YYSYMBOL_PROBES = 69, /* PROBES */ + YYSYMBOL_IVFFLAT = 70, /* IVFFLAT */ + YYSYMBOL_EQ = 71, /* EQ */ + YYSYMBOL_LT = 72, /* LT */ + YYSYMBOL_GT = 73, /* GT */ + YYSYMBOL_LE = 74, /* LE */ + YYSYMBOL_GE = 75, /* GE */ + YYSYMBOL_NE = 76, /* NE */ + YYSYMBOL_LIKE = 77, /* LIKE */ + YYSYMBOL_IS = 78, /* IS */ + YYSYMBOL_NUMBER = 79, /* NUMBER */ + YYSYMBOL_FLOAT = 80, /* FLOAT */ + YYSYMBOL_ID = 81, /* ID */ + YYSYMBOL_SSS = 82, /* SSS */ + YYSYMBOL_83_ = 83, /* '+' */ + YYSYMBOL_84_ = 84, /* '-' */ + YYSYMBOL_85_ = 85, /* '*' */ + YYSYMBOL_86_ = 86, /* '/' */ + YYSYMBOL_UMINUS = 87, /* UMINUS */ + YYSYMBOL_YYACCEPT = 88, /* $accept */ + YYSYMBOL_commands = 89, /* commands */ + YYSYMBOL_command_wrapper = 90, /* command_wrapper */ + YYSYMBOL_exit_stmt = 91, /* exit_stmt */ + YYSYMBOL_help_stmt = 92, /* help_stmt */ + YYSYMBOL_sync_stmt = 93, /* sync_stmt */ + YYSYMBOL_begin_stmt = 94, /* begin_stmt */ + YYSYMBOL_commit_stmt = 95, /* commit_stmt */ + YYSYMBOL_rollback_stmt = 96, /* rollback_stmt */ + YYSYMBOL_drop_table_stmt = 97, /* drop_table_stmt */ + YYSYMBOL_show_tables_stmt = 98, /* show_tables_stmt */ + YYSYMBOL_desc_table_stmt = 99, /* desc_table_stmt */ + YYSYMBOL_show_index_stmt = 100, /* show_index_stmt */ + YYSYMBOL_create_index_stmt = 101, /* create_index_stmt */ + YYSYMBOL_opt_unique = 102, /* opt_unique */ + YYSYMBOL_index_type = 103, /* index_type */ + YYSYMBOL_vector_index_config = 104, /* vector_index_config */ + YYSYMBOL_attr_list = 105, /* attr_list */ + YYSYMBOL_drop_index_stmt = 106, /* drop_index_stmt */ + YYSYMBOL_create_table_stmt = 107, /* create_table_stmt */ + YYSYMBOL_create_view_stmt = 108, /* create_view_stmt */ + YYSYMBOL_drop_view_stmt = 109, /* drop_view_stmt */ + YYSYMBOL_attr_def_list = 110, /* attr_def_list */ + YYSYMBOL_attr_def = 111, /* attr_def */ + YYSYMBOL_nullable_constraint = 112, /* nullable_constraint */ + YYSYMBOL_type = 113, /* type */ + YYSYMBOL_insert_stmt = 114, /* insert_stmt */ + YYSYMBOL_values_list = 115, /* values_list */ + YYSYMBOL_digits = 116, /* digits */ + YYSYMBOL_digits_list = 117, /* digits_list */ + YYSYMBOL_value_list = 118, /* value_list */ + YYSYMBOL_value = 119, /* value */ + YYSYMBOL_nonnegative_value = 120, /* nonnegative_value */ + YYSYMBOL_storage_format = 121, /* storage_format */ + YYSYMBOL_delete_stmt = 122, /* delete_stmt */ + YYSYMBOL_update_stmt = 123, /* update_stmt */ + YYSYMBOL_set_clauses = 124, /* set_clauses */ + YYSYMBOL_set_clause = 125, /* set_clause */ + YYSYMBOL_select_stmt = 126, /* select_stmt */ + YYSYMBOL_calc_stmt = 127, /* calc_stmt */ + YYSYMBOL_expression_list = 128, /* expression_list */ + YYSYMBOL_expression = 129, /* expression */ + YYSYMBOL_alias = 130, /* alias */ + YYSYMBOL_func_expr = 131, /* func_expr */ + YYSYMBOL_sub_query_expr = 132, /* sub_query_expr */ + YYSYMBOL_rel_attr = 133, /* rel_attr */ + YYSYMBOL_relation = 134, /* relation */ + YYSYMBOL_rel_list = 135, /* rel_list */ + YYSYMBOL_join_clauses = 136, /* join_clauses */ + YYSYMBOL_where = 137, /* where */ + YYSYMBOL_condition = 138, /* condition */ + YYSYMBOL_comp_op = 139, /* comp_op */ + YYSYMBOL_opt_order_by = 140, /* opt_order_by */ + YYSYMBOL_sort_list = 141, /* sort_list */ + YYSYMBOL_sort_unit = 142, /* sort_unit */ + YYSYMBOL_group_by = 143, /* group_by */ + YYSYMBOL_opt_having = 144, /* opt_having */ + YYSYMBOL_opt_limit = 145, /* opt_limit */ + YYSYMBOL_explain_stmt = 146, /* explain_stmt */ + YYSYMBOL_set_variable_stmt = 147, /* set_variable_stmt */ + YYSYMBOL_opt_semicolon = 148 /* opt_semicolon */ }; typedef enum yysymbol_kind_t yysymbol_kind_t; - - - #ifdef short -# undef short +#undef short #endif /* On compilers that do not define __PTRDIFF_MAX__ etc., make sure @@ -265,11 +328,11 @@ typedef enum yysymbol_kind_t yysymbol_kind_t; so that the code can choose integer types of a good width. */ #ifndef __PTRDIFF_MAX__ -# include /* INFRINGES ON USER NAME SPACE */ -# if defined __STDC_VERSION__ && 199901 <= __STDC_VERSION__ -# include /* INFRINGES ON USER NAME SPACE */ -# define YY_STDINT_H -# endif +#include /* INFRINGES ON USER NAME SPACE */ +#if defined __STDC_VERSION__ && 199901 <= __STDC_VERSION__ +#include /* INFRINGES ON USER NAME SPACE */ +#define YY_STDINT_H +#endif #endif /* Narrow types that promote to a signed type and that can represent a @@ -299,16 +362,15 @@ typedef short yytype_int16; (aka HP-UX 11i v2) only through the end of 2022; see Table 2 of . */ #ifdef __hpux -# undef UINT_LEAST8_MAX -# undef UINT_LEAST16_MAX -# define UINT_LEAST8_MAX 255 -# define UINT_LEAST16_MAX 65535 +#undef UINT_LEAST8_MAX +#undef UINT_LEAST16_MAX +#define UINT_LEAST8_MAX 255 +#define UINT_LEAST16_MAX 65535 #endif #if defined __UINT_LEAST8_MAX__ && __UINT_LEAST8_MAX__ <= __INT_MAX__ typedef __UINT_LEAST8_TYPE__ yytype_uint8; -#elif (!defined __UINT_LEAST8_MAX__ && defined YY_STDINT_H \ - && UINT_LEAST8_MAX <= INT_MAX) +#elif (!defined __UINT_LEAST8_MAX__ && defined YY_STDINT_H && UINT_LEAST8_MAX <= INT_MAX) typedef uint_least8_t yytype_uint8; #elif !defined __UINT_LEAST8_MAX__ && UCHAR_MAX <= INT_MAX typedef unsigned char yytype_uint8; @@ -318,8 +380,7 @@ typedef short yytype_uint8; #if defined __UINT_LEAST16_MAX__ && __UINT_LEAST16_MAX__ <= __INT_MAX__ typedef __UINT_LEAST16_TYPE__ yytype_uint16; -#elif (!defined __UINT_LEAST16_MAX__ && defined YY_STDINT_H \ - && UINT_LEAST16_MAX <= INT_MAX) +#elif (!defined __UINT_LEAST16_MAX__ && defined YY_STDINT_H && UINT_LEAST16_MAX <= INT_MAX) typedef uint_least16_t yytype_uint16; #elif !defined __UINT_LEAST16_MAX__ && USHRT_MAX <= INT_MAX typedef unsigned short yytype_uint16; @@ -328,573 +389,3109 @@ typedef int yytype_uint16; #endif #ifndef YYPTRDIFF_T -# if defined __PTRDIFF_TYPE__ && defined __PTRDIFF_MAX__ -# define YYPTRDIFF_T __PTRDIFF_TYPE__ -# define YYPTRDIFF_MAXIMUM __PTRDIFF_MAX__ -# elif defined PTRDIFF_MAX -# ifndef ptrdiff_t -# include /* INFRINGES ON USER NAME SPACE */ -# endif -# define YYPTRDIFF_T ptrdiff_t -# define YYPTRDIFF_MAXIMUM PTRDIFF_MAX -# else -# define YYPTRDIFF_T long -# define YYPTRDIFF_MAXIMUM LONG_MAX -# endif +#if defined __PTRDIFF_TYPE__ && defined __PTRDIFF_MAX__ +#define YYPTRDIFF_T __PTRDIFF_TYPE__ +#define YYPTRDIFF_MAXIMUM __PTRDIFF_MAX__ +#elif defined PTRDIFF_MAX +#ifndef ptrdiff_t +#include /* INFRINGES ON USER NAME SPACE */ +#endif +#define YYPTRDIFF_T ptrdiff_t +#define YYPTRDIFF_MAXIMUM PTRDIFF_MAX +#else +#define YYPTRDIFF_T long +#define YYPTRDIFF_MAXIMUM LONG_MAX +#endif #endif #ifndef YYSIZE_T -# ifdef __SIZE_TYPE__ -# define YYSIZE_T __SIZE_TYPE__ -# elif defined size_t -# define YYSIZE_T size_t -# elif defined __STDC_VERSION__ && 199901 <= __STDC_VERSION__ -# include /* INFRINGES ON USER NAME SPACE */ -# define YYSIZE_T size_t -# else -# define YYSIZE_T unsigned -# endif +#ifdef __SIZE_TYPE__ +#define YYSIZE_T __SIZE_TYPE__ +#elif defined size_t +#define YYSIZE_T size_t +#elif defined __STDC_VERSION__ && 199901 <= __STDC_VERSION__ +#include /* INFRINGES ON USER NAME SPACE */ +#define YYSIZE_T size_t +#else +#define YYSIZE_T unsigned +#endif #endif -#define YYSIZE_MAXIMUM \ - YY_CAST (YYPTRDIFF_T, \ - (YYPTRDIFF_MAXIMUM < YY_CAST (YYSIZE_T, -1) \ - ? YYPTRDIFF_MAXIMUM \ - : YY_CAST (YYSIZE_T, -1))) - -#define YYSIZEOF(X) YY_CAST (YYPTRDIFF_T, sizeof (X)) +#define YYSIZE_MAXIMUM \ + YY_CAST(YYPTRDIFF_T, (YYPTRDIFF_MAXIMUM < YY_CAST(YYSIZE_T, -1) ? YYPTRDIFF_MAXIMUM : YY_CAST(YYSIZE_T, -1))) +#define YYSIZEOF(X) YY_CAST(YYPTRDIFF_T, sizeof(X)) /* Stored state numbers (used for stacks). */ -typedef yytype_uint8 yy_state_t; +typedef yytype_int16 yy_state_t; /* State numbers in computations. */ typedef int yy_state_fast_t; #ifndef YY_ -# if defined YYENABLE_NLS && YYENABLE_NLS -# if ENABLE_NLS -# include /* INFRINGES ON USER NAME SPACE */ -# define YY_(Msgid) dgettext ("bison-runtime", Msgid) -# endif -# endif -# ifndef YY_ -# define YY_(Msgid) Msgid -# endif +#if defined YYENABLE_NLS && YYENABLE_NLS +#if ENABLE_NLS +#include /* INFRINGES ON USER NAME SPACE */ +#define YY_(Msgid) dgettext("bison-runtime", Msgid) +#endif +#endif +#ifndef YY_ +#define YY_(Msgid) Msgid +#endif #endif - #ifndef YY_ATTRIBUTE_PURE -# if defined __GNUC__ && 2 < __GNUC__ + (96 <= __GNUC_MINOR__) -# define YY_ATTRIBUTE_PURE __attribute__ ((__pure__)) -# else -# define YY_ATTRIBUTE_PURE -# endif +#if defined __GNUC__ && 2 < __GNUC__ + (96 <= __GNUC_MINOR__) +#define YY_ATTRIBUTE_PURE __attribute__((__pure__)) +#else +#define YY_ATTRIBUTE_PURE +#endif #endif #ifndef YY_ATTRIBUTE_UNUSED -# if defined __GNUC__ && 2 < __GNUC__ + (7 <= __GNUC_MINOR__) -# define YY_ATTRIBUTE_UNUSED __attribute__ ((__unused__)) -# else -# define YY_ATTRIBUTE_UNUSED -# endif +#if defined __GNUC__ && 2 < __GNUC__ + (7 <= __GNUC_MINOR__) +#define YY_ATTRIBUTE_UNUSED __attribute__((__unused__)) +#else +#define YY_ATTRIBUTE_UNUSED +#endif #endif /* Suppress unused-variable warnings by "using" E. */ -#if ! defined lint || defined __GNUC__ -# define YY_USE(E) ((void) (E)) +#if !defined lint || defined __GNUC__ +#define YY_USE(E) ((void)(E)) #else -# define YY_USE(E) /* empty */ +#define YY_USE(E) /* empty */ #endif /* Suppress an incorrect diagnostic about yylval being uninitialized. */ -#if defined __GNUC__ && ! defined __ICC && 406 <= __GNUC__ * 100 + __GNUC_MINOR__ -# if __GNUC__ * 100 + __GNUC_MINOR__ < 407 -# define YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN \ - _Pragma ("GCC diagnostic push") \ - _Pragma ("GCC diagnostic ignored \"-Wuninitialized\"") -# else -# define YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN \ - _Pragma ("GCC diagnostic push") \ - _Pragma ("GCC diagnostic ignored \"-Wuninitialized\"") \ - _Pragma ("GCC diagnostic ignored \"-Wmaybe-uninitialized\"") -# endif -# define YY_IGNORE_MAYBE_UNINITIALIZED_END \ - _Pragma ("GCC diagnostic pop") +#if defined __GNUC__ && !defined __ICC && 406 <= __GNUC__ * 100 + __GNUC_MINOR__ +#if __GNUC__ * 100 + __GNUC_MINOR__ < 407 +#define YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN \ + _Pragma("GCC diagnostic push") _Pragma("GCC diagnostic ignored \"-Wuninitialized\"") +#else +#define YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN \ + _Pragma("GCC diagnostic push") _Pragma("GCC diagnostic ignored \"-Wuninitialized\"") \ + _Pragma("GCC diagnostic ignored \"-Wmaybe-uninitialized\"") +#endif +#define YY_IGNORE_MAYBE_UNINITIALIZED_END _Pragma("GCC diagnostic pop") #else -# define YY_INITIAL_VALUE(Value) Value +#define YY_INITIAL_VALUE(Value) Value #endif #ifndef YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN -# define YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN -# define YY_IGNORE_MAYBE_UNINITIALIZED_END +#define YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN +#define YY_IGNORE_MAYBE_UNINITIALIZED_END #endif #ifndef YY_INITIAL_VALUE -# define YY_INITIAL_VALUE(Value) /* Nothing. */ +#define YY_INITIAL_VALUE(Value) /* Nothing. */ #endif -#if defined __cplusplus && defined __GNUC__ && ! defined __ICC && 6 <= __GNUC__ -# define YY_IGNORE_USELESS_CAST_BEGIN \ - _Pragma ("GCC diagnostic push") \ - _Pragma ("GCC diagnostic ignored \"-Wuseless-cast\"") -# define YY_IGNORE_USELESS_CAST_END \ - _Pragma ("GCC diagnostic pop") +#if defined __cplusplus && defined __GNUC__ && !defined __ICC && 6 <= __GNUC__ +#define YY_IGNORE_USELESS_CAST_BEGIN _Pragma("GCC diagnostic push") _Pragma("GCC diagnostic ignored \"-Wuseless-cast\"") +#define YY_IGNORE_USELESS_CAST_END _Pragma("GCC diagnostic pop") #endif #ifndef YY_IGNORE_USELESS_CAST_BEGIN -# define YY_IGNORE_USELESS_CAST_BEGIN -# define YY_IGNORE_USELESS_CAST_END +#define YY_IGNORE_USELESS_CAST_BEGIN +#define YY_IGNORE_USELESS_CAST_END #endif - -#define YY_ASSERT(E) ((void) (0 && (E))) +#define YY_ASSERT(E) ((void)(0 && (E))) #if 1 /* The parser invokes alloca or malloc; define the necessary symbols. */ -# ifdef YYSTACK_USE_ALLOCA -# if YYSTACK_USE_ALLOCA -# ifdef __GNUC__ -# define YYSTACK_ALLOC __builtin_alloca -# elif defined __BUILTIN_VA_ARG_INCR -# include /* INFRINGES ON USER NAME SPACE */ -# elif defined _AIX -# define YYSTACK_ALLOC __alloca -# elif defined _MSC_VER -# include /* INFRINGES ON USER NAME SPACE */ -# define alloca _alloca -# else -# define YYSTACK_ALLOC alloca -# if ! defined _ALLOCA_H && ! defined EXIT_SUCCESS -# include /* INFRINGES ON USER NAME SPACE */ - /* Use EXIT_SUCCESS as a witness for stdlib.h. */ -# ifndef EXIT_SUCCESS -# define EXIT_SUCCESS 0 -# endif -# endif -# endif -# endif -# endif - -# ifdef YYSTACK_ALLOC - /* Pacify GCC's 'empty if-body' warning. */ -# define YYSTACK_FREE(Ptr) do { /* empty */; } while (0) -# ifndef YYSTACK_ALLOC_MAXIMUM - /* The OS might guarantee only one guard page at the bottom of the stack, - and a page size can be as small as 4096 bytes. So we cannot safely - invoke alloca (N) if N exceeds 4096. Use a slightly smaller number - to allow for a few compiler-allocated temporary stack slots. */ -# define YYSTACK_ALLOC_MAXIMUM 4032 /* reasonable circa 2006 */ -# endif -# else -# define YYSTACK_ALLOC YYMALLOC -# define YYSTACK_FREE YYFREE -# ifndef YYSTACK_ALLOC_MAXIMUM -# define YYSTACK_ALLOC_MAXIMUM YYSIZE_MAXIMUM -# endif -# if (defined __cplusplus && ! defined EXIT_SUCCESS \ - && ! ((defined YYMALLOC || defined malloc) \ - && (defined YYFREE || defined free))) -# include /* INFRINGES ON USER NAME SPACE */ -# ifndef EXIT_SUCCESS -# define EXIT_SUCCESS 0 -# endif -# endif -# ifndef YYMALLOC -# define YYMALLOC malloc -# if ! defined malloc && ! defined EXIT_SUCCESS -void *malloc (YYSIZE_T); /* INFRINGES ON USER NAME SPACE */ -# endif -# endif -# ifndef YYFREE -# define YYFREE free -# if ! defined free && ! defined EXIT_SUCCESS -void free (void *); /* INFRINGES ON USER NAME SPACE */ -# endif -# endif -# endif +#ifdef YYSTACK_USE_ALLOCA +#if YYSTACK_USE_ALLOCA +#ifdef __GNUC__ +#define YYSTACK_ALLOC __builtin_alloca +#elif defined __BUILTIN_VA_ARG_INCR +#include /* INFRINGES ON USER NAME SPACE */ +#elif defined _AIX +#define YYSTACK_ALLOC __alloca +#elif defined _MSC_VER +#include /* INFRINGES ON USER NAME SPACE */ +#define alloca _alloca +#else +#define YYSTACK_ALLOC alloca +#if !defined _ALLOCA_H && !defined EXIT_SUCCESS +#include /* INFRINGES ON USER NAME SPACE */ +/* Use EXIT_SUCCESS as a witness for stdlib.h. */ +#ifndef EXIT_SUCCESS +#define EXIT_SUCCESS 0 +#endif +#endif +#endif +#endif +#endif + +#ifdef YYSTACK_ALLOC +/* Pacify GCC's 'empty if-body' warning. */ +#define YYSTACK_FREE(Ptr) \ + do { /* empty */ \ + ; \ + } while (0) +#ifndef YYSTACK_ALLOC_MAXIMUM +/* The OS might guarantee only one guard page at the bottom of the stack, + and a page size can be as small as 4096 bytes. So we cannot safely + invoke alloca (N) if N exceeds 4096. Use a slightly smaller number + to allow for a few compiler-allocated temporary stack slots. */ +#define YYSTACK_ALLOC_MAXIMUM 4032 /* reasonable circa 2006 */ +#endif +#else +#define YYSTACK_ALLOC YYMALLOC +#define YYSTACK_FREE YYFREE +#ifndef YYSTACK_ALLOC_MAXIMUM +#define YYSTACK_ALLOC_MAXIMUM YYSIZE_MAXIMUM +#endif +#if (defined __cplusplus && !defined EXIT_SUCCESS && \ + !((defined YYMALLOC || defined malloc) && (defined YYFREE || defined free))) +#include /* INFRINGES ON USER NAME SPACE */ +#ifndef EXIT_SUCCESS +#define EXIT_SUCCESS 0 +#endif +#endif +#ifndef YYMALLOC +#define YYMALLOC malloc +#if !defined malloc && !defined EXIT_SUCCESS +void *malloc(YYSIZE_T); /* INFRINGES ON USER NAME SPACE */ +#endif +#endif +#ifndef YYFREE +#define YYFREE free +#if !defined free && !defined EXIT_SUCCESS +void free(void *); /* INFRINGES ON USER NAME SPACE */ +#endif +#endif +#endif #endif /* 1 */ -#if (! defined yyoverflow \ - && (! defined __cplusplus \ - || (defined YYLTYPE_IS_TRIVIAL && YYLTYPE_IS_TRIVIAL \ - && defined YYSTYPE_IS_TRIVIAL && YYSTYPE_IS_TRIVIAL))) +#if (!defined yyoverflow && (!defined __cplusplus || (defined YYLTYPE_IS_TRIVIAL && YYLTYPE_IS_TRIVIAL && \ + defined YYSTYPE_IS_TRIVIAL && YYSTYPE_IS_TRIVIAL))) /* A type that is properly aligned for any stack member. */ union yyalloc { yy_state_t yyss_alloc; - YYSTYPE yyvs_alloc; - YYLTYPE yyls_alloc; + YYSTYPE yyvs_alloc; + YYLTYPE yyls_alloc; }; /* The size of the maximum gap between one aligned stack and the next. */ -# define YYSTACK_GAP_MAXIMUM (YYSIZEOF (union yyalloc) - 1) +#define YYSTACK_GAP_MAXIMUM (YYSIZEOF(union yyalloc) - 1) /* The size of an array large to enough to hold all stacks, each with N elements. */ -# define YYSTACK_BYTES(N) \ - ((N) * (YYSIZEOF (yy_state_t) + YYSIZEOF (YYSTYPE) \ - + YYSIZEOF (YYLTYPE)) \ - + 2 * YYSTACK_GAP_MAXIMUM) +#define YYSTACK_BYTES(N) \ + ((N) * (YYSIZEOF(yy_state_t) + YYSIZEOF(YYSTYPE) + YYSIZEOF(YYLTYPE)) + 2 * YYSTACK_GAP_MAXIMUM) -# define YYCOPY_NEEDED 1 +#define YYCOPY_NEEDED 1 /* Relocate STACK from its old location to the new one. The local variables YYSIZE and YYSTACKSIZE give the old and new number of elements in the stack, and YYPTR gives the new location of the stack. Advance YYPTR to a properly aligned location for the next stack. */ -# define YYSTACK_RELOCATE(Stack_alloc, Stack) \ - do \ - { \ - YYPTRDIFF_T yynewbytes; \ - YYCOPY (&yyptr->Stack_alloc, Stack, yysize); \ - Stack = &yyptr->Stack_alloc; \ - yynewbytes = yystacksize * YYSIZEOF (*Stack) + YYSTACK_GAP_MAXIMUM; \ - yyptr += yynewbytes / YYSIZEOF (*yyptr); \ - } \ - while (0) +#define YYSTACK_RELOCATE(Stack_alloc, Stack) \ + do { \ + YYPTRDIFF_T yynewbytes; \ + YYCOPY(&yyptr->Stack_alloc, Stack, yysize); \ + Stack = &yyptr->Stack_alloc; \ + yynewbytes = yystacksize * YYSIZEOF(*Stack) + YYSTACK_GAP_MAXIMUM; \ + yyptr += yynewbytes / YYSIZEOF(*yyptr); \ + } while (0) #endif #if defined YYCOPY_NEEDED && YYCOPY_NEEDED /* Copy COUNT objects from SRC to DST. The source and destination do not overlap. */ -# ifndef YYCOPY -# if defined __GNUC__ && 1 < __GNUC__ -# define YYCOPY(Dst, Src, Count) \ - __builtin_memcpy (Dst, Src, YY_CAST (YYSIZE_T, (Count)) * sizeof (*(Src))) -# else -# define YYCOPY(Dst, Src, Count) \ - do \ - { \ - YYPTRDIFF_T yyi; \ - for (yyi = 0; yyi < (Count); yyi++) \ - (Dst)[yyi] = (Src)[yyi]; \ - } \ - while (0) -# endif -# endif +#ifndef YYCOPY +#if defined __GNUC__ && 1 < __GNUC__ +#define YYCOPY(Dst, Src, Count) __builtin_memcpy(Dst, Src, YY_CAST(YYSIZE_T, (Count)) * sizeof(*(Src))) +#else +#define YYCOPY(Dst, Src, Count) \ + do { \ + YYPTRDIFF_T yyi; \ + for (yyi = 0; yyi < (Count); yyi++) \ + (Dst)[yyi] = (Src)[yyi]; \ + } while (0) +#endif +#endif #endif /* !YYCOPY_NEEDED */ /* YYFINAL -- State number of the termination state. */ -#define YYFINAL 65 +#define YYFINAL 76 /* YYLAST -- Last index in YYTABLE. */ -#define YYLAST 141 +#define YYLAST 346 /* YYNTOKENS -- Number of terminals. */ -#define YYNTOKENS 60 +#define YYNTOKENS 88 /* YYNNTS -- Number of nonterminals. */ -#define YYNNTS 41 +#define YYNNTS 61 /* YYNRULES -- Number of rules. */ -#define YYNRULES 92 +#define YYNRULES 165 /* YYNSTATES -- Number of states. */ -#define YYNSTATES 166 +#define YYNSTATES 323 /* YYMAXUTOK -- Last valid token kind. */ -#define YYMAXUTOK 310 - +#define YYMAXUTOK 338 /* YYTRANSLATE(TOKEN-NUM) -- Symbol number corresponding to TOKEN-NUM as returned by yylex, with out-of-bounds checking. */ -#define YYTRANSLATE(YYX) \ - (0 <= (YYX) && (YYX) <= YYMAXUTOK \ - ? YY_CAST (yysymbol_kind_t, yytranslate[YYX]) \ - : YYSYMBOL_YYUNDEF) +#define YYTRANSLATE(YYX) \ + (0 <= (YYX) && (YYX) <= YYMAXUTOK ? YY_CAST(yysymbol_kind_t, yytranslate[YYX]) : YYSYMBOL_YYUNDEF) /* YYTRANSLATE[TOKEN-NUM] -- Symbol number corresponding to TOKEN-NUM as returned by yylex. */ -static const yytype_int8 yytranslate[] = -{ - 0, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 57, 55, 2, 56, 2, 58, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 2, 2, 2, 2, - 2, 2, 2, 2, 2, 2, 1, 2, 3, 4, - 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, - 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, - 25, 26, 27, 28, 29, 30, 31, 32, 33, 34, - 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, - 45, 46, 47, 48, 49, 50, 51, 52, 53, 54, - 59 -}; +static const yytype_int8 yytranslate[] = {0, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 85, + 83, + 2, + 84, + 2, + 86, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 2, + 1, + 2, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 11, + 12, + 13, + 14, + 15, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 31, + 32, + 33, + 34, + 35, + 36, + 37, + 38, + 39, + 40, + 41, + 42, + 43, + 44, + 45, + 46, + 47, + 48, + 49, + 50, + 51, + 52, + 53, + 54, + 55, + 56, + 57, + 58, + 59, + 60, + 61, + 62, + 63, + 64, + 65, + 66, + 67, + 68, + 69, + 70, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 87}; #if YYDEBUG /* YYRLINE[YYN] -- Source line where rule number YYN was defined. */ -static const yytype_int16 yyrline[] = -{ - 0, 189, 189, 197, 198, 199, 200, 201, 202, 203, - 204, 205, 206, 207, 208, 209, 210, 211, 212, 213, - 214, 215, 216, 220, 226, 231, 237, 243, 249, 255, - 262, 268, 276, 290, 300, 324, 327, 340, 348, 358, - 361, 362, 363, 364, 367, 384, 387, 398, 402, 406, - 415, 418, 425, 437, 452, 477, 486, 491, 502, 505, - 508, 511, 514, 518, 521, 526, 532, 539, 544, 554, - 559, 564, 578, 581, 587, 590, 595, 602, 614, 626, - 638, 653, 654, 655, 656, 657, 658, 664, 669, 682, - 690, 700, 701 -}; +static const yytype_int16 yyrline[] = {0, + 282, + 282, + 290, + 291, + 292, + 293, + 294, + 295, + 296, + 297, + 298, + 299, + 300, + 301, + 302, + 303, + 304, + 305, + 306, + 307, + 308, + 309, + 310, + 311, + 315, + 321, + 326, + 332, + 338, + 344, + 350, + 357, + 363, + 371, + 381, + 393, + 409, + 410, + 414, + 421, + 428, + 437, + 449, + 455, + 464, + 474, + 478, + 482, + 486, + 490, + 497, + 505, + 517, + 527, + 530, + 543, + 561, + 590, + 594, + 598, + 603, + 609, + 610, + 611, + 612, + 613, + 614, + 618, + 628, + 642, + 648, + 655, + 659, + 663, + 667, + 675, + 678, + 683, + 691, + 694, + 700, + 708, + 711, + 715, + 722, + 726, + 730, + 736, + 739, + 742, + 745, + 752, + 755, + 762, + 774, + 788, + 793, + 800, + 810, + 848, + 881, + 887, + 896, + 899, + 908, + 924, + 927, + 930, + 933, + 936, + 944, + 947, + 952, + 958, + 961, + 964, + 967, + 974, + 977, + 980, + 985, + 993, + 1000, + 1005, + 1015, + 1021, + 1031, + 1048, + 1055, + 1067, + 1070, + 1076, + 1080, + 1087, + 1091, + 1098, + 1099, + 1100, + 1101, + 1102, + 1103, + 1104, + 1105, + 1106, + 1107, + 1108, + 1109, + 1110, + 1111, + 1116, + 1119, + 1127, + 1132, + 1140, + 1146, + 1152, + 1162, + 1165, + 1173, + 1176, + 1184, + 1187, + 1195, + 1203, + 1214}; #endif /** Accessing symbol of state STATE. */ -#define YY_ACCESSING_SYMBOL(State) YY_CAST (yysymbol_kind_t, yystos[State]) +#define YY_ACCESSING_SYMBOL(State) YY_CAST(yysymbol_kind_t, yystos[State]) #if 1 /* The user-facing name of the symbol whose (internal) number is YYSYMBOL. No bounds checking. */ -static const char *yysymbol_name (yysymbol_kind_t yysymbol) YY_ATTRIBUTE_UNUSED; +static const char *yysymbol_name(yysymbol_kind_t yysymbol) YY_ATTRIBUTE_UNUSED; /* YYTNAME[SYMBOL-NUM] -- String name of the symbol SYMBOL-NUM. First, the terminals, then, starting at YYNTOKENS, nonterminals. */ -static const char *const yytname[] = -{ - "\"end of file\"", "error", "\"invalid token\"", "SEMICOLON", "BY", - "CREATE", "DROP", "GROUP", "TABLE", "TABLES", "INDEX", "CALC", "SELECT", - "DESC", "SHOW", "SYNC", "INSERT", "DELETE", "UPDATE", "LBRACE", "RBRACE", - "COMMA", "TRX_BEGIN", "TRX_COMMIT", "TRX_ROLLBACK", "INT_T", "STRING_T", - "FLOAT_T", "VECTOR_T", "HELP", "EXIT", "DOT", "INTO", "VALUES", "FROM", - "WHERE", "AND", "SET", "ON", "LOAD", "DATA", "INFILE", "EXPLAIN", - "STORAGE", "FORMAT", "EQ", "LT", "GT", "LE", "GE", "NE", "NUMBER", - "FLOAT", "ID", "SSS", "'+'", "'-'", "'*'", "'/'", "UMINUS", "$accept", - "commands", "command_wrapper", "exit_stmt", "help_stmt", "sync_stmt", - "begin_stmt", "commit_stmt", "rollback_stmt", "drop_table_stmt", - "show_tables_stmt", "desc_table_stmt", "create_index_stmt", - "drop_index_stmt", "create_table_stmt", "attr_def_list", "attr_def", - "number", "type", "insert_stmt", "value_list", "value", "storage_format", - "delete_stmt", "update_stmt", "select_stmt", "calc_stmt", - "expression_list", "expression", "rel_attr", "relation", "rel_list", - "where", "condition_list", "condition", "comp_op", "group_by", - "load_data_stmt", "explain_stmt", "set_variable_stmt", "opt_semicolon", YY_NULLPTR -}; - -static const char * -yysymbol_name (yysymbol_kind_t yysymbol) -{ - return yytname[yysymbol]; -} +static const char *const yytname[] = {"\"end of file\"", + "error", + "\"invalid token\"", + "SEMICOLON", + "AS", + "ASC", + "BY", + "CREATE", + "DROP", + "EXISTS", + "GROUP", + "HAVING", + "ORDER", + "TABLE", + "TABLES", + "INDEX", + "CALC", + "SELECT", + "DESC", + "SHOW", + "SYNC", + "INSERT", + "DELETE", + "UPDATE", + "LBRACE", + "RBRACE", + "LSBRACE", + "RSBRACE", + "COMMA", + "TRX_BEGIN", + "TRX_COMMIT", + "TRX_ROLLBACK", + "INT_T", + "IN", + "TRUE", + "FALSE", + "STRING_T", + "FLOAT_T", + "DATE_T", + "TEXT_T", + "VECTOR_T", + "NOT", + "UNIQUE", + "NULL_T", + "LIMIT", + "NULLABLE", + "HELP", + "QUOTE", + "EXIT", + "DOT", + "INTO", + "VALUES", + "FROM", + "WHERE", + "AND", + "OR", + "SET", + "ON", + "INFILE", + "EXPLAIN", + "STORAGE", + "FORMAT", + "INNER", + "JOIN", + "VIEW", + "WITH", + "DISTANCE", + "TYPE", + "LISTS", + "PROBES", + "IVFFLAT", + "EQ", + "LT", + "GT", + "LE", + "GE", + "NE", + "LIKE", + "IS", + "NUMBER", + "FLOAT", + "ID", + "SSS", + "'+'", + "'-'", + "'*'", + "'/'", + "UMINUS", + "$accept", + "commands", + "command_wrapper", + "exit_stmt", + "help_stmt", + "sync_stmt", + "begin_stmt", + "commit_stmt", + "rollback_stmt", + "drop_table_stmt", + "show_tables_stmt", + "desc_table_stmt", + "show_index_stmt", + "create_index_stmt", + "opt_unique", + "index_type", + "vector_index_config", + "attr_list", + "drop_index_stmt", + "create_table_stmt", + "create_view_stmt", + "drop_view_stmt", + "attr_def_list", + "attr_def", + "nullable_constraint", + "type", + "insert_stmt", + "values_list", + "digits", + "digits_list", + "value_list", + "value", + "nonnegative_value", + "storage_format", + "delete_stmt", + "update_stmt", + "set_clauses", + "set_clause", + "select_stmt", + "calc_stmt", + "expression_list", + "expression", + "alias", + "func_expr", + "sub_query_expr", + "rel_attr", + "relation", + "rel_list", + "join_clauses", + "where", + "condition", + "comp_op", + "opt_order_by", + "sort_list", + "sort_unit", + "group_by", + "opt_having", + "opt_limit", + "explain_stmt", + "set_variable_stmt", + "opt_semicolon", + YY_NULLPTR}; + +static const char *yysymbol_name(yysymbol_kind_t yysymbol) { return yytname[yysymbol]; } #endif -#define YYPACT_NINF (-97) +#define YYPACT_NINF (-203) -#define yypact_value_is_default(Yyn) \ - ((Yyn) == YYPACT_NINF) +#define yypact_value_is_default(Yyn) ((Yyn) == YYPACT_NINF) #define YYTABLE_NINF (-1) -#define yytable_value_is_error(Yyn) \ - 0 +#define yytable_value_is_error(Yyn) 0 /* YYPACT[STATE-NUM] -- Index in YYTABLE of the portion describing STATE-NUM. */ -static const yytype_int8 yypact[] = -{ - 51, 6, 14, -16, -16, -38, 2, -97, -5, -3, - -24, -97, -97, -97, -97, -97, -23, -6, 51, 32, - 36, -97, -97, -97, -97, -97, -97, -97, -97, -97, - -97, -97, -97, -97, -97, -97, -97, -97, -97, -97, - -97, -2, 17, 23, 25, -16, -97, -97, 24, -97, - -16, -97, -97, -97, -8, -97, 45, -97, -97, 38, - 39, 52, 49, 54, -97, -97, -97, -97, 77, 59, - -97, 60, -12, 46, -97, -16, -16, -16, -16, -16, - 47, 68, 67, 55, -42, 53, 56, 57, 58, -97, - -97, -97, -32, -32, -97, -97, -97, 91, 67, 94, - -47, -97, 69, -97, 83, -7, 95, 98, -97, 47, - -97, -42, 37, 37, -97, 82, -42, 111, -97, -97, - -97, -97, 101, 56, 102, 70, -97, -97, 100, -97, - -97, -97, -97, -97, -97, -47, -47, -47, 67, 71, - 74, 95, 84, 106, -42, 108, -97, -97, -97, -97, - -97, -97, -97, -97, 109, -97, 86, -97, -97, 100, - -97, -97, 87, -97, 78, -97 -}; +static const yytype_int16 yypact[] = {281, + 13, + 4, + 140, + 140, + -66, + 46, + -203, + -14, + 51, + -43, + -203, + -203, + -203, + -203, + -203, + -34, + 281, + 85, + 106, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + 53, + 113, + -203, + 61, + 123, + 75, + 84, + 88, + 32, + 47, + -203, + -203, + -203, + -203, + -203, + -8, + -203, + 140, + -203, + -203, + -203, + 8, + -203, + -203, + -203, + 95, + -203, + -203, + 118, + 90, + 91, + 117, + 108, + -203, + -203, + -203, + -203, + -6, + 96, + 26, + 97, + -203, + 119, + -203, + 140, + 157, + 159, + -203, + -203, + 57, + -203, + -5, + 140, + -56, + -203, + 104, + -203, + 140, + 140, + 140, + 140, + 161, + 109, + 109, + -3, + 134, + 110, + 160, + 111, + 127, + 16, + 136, + 180, + 120, + 141, + 121, + 95, + -203, + -203, + -203, + -203, + -203, + 47, + 174, + -203, + -203, + -203, + 58, + 58, + -203, + -203, + 140, + -203, + 0, + 134, + -203, + 120, + 181, + 202, + -203, + 133, + 6, + -203, + 66, + -203, + -203, + 83, + 178, + 142, + 180, + -203, + 146, + -203, + 195, + 205, + 150, + -203, + -203, + -203, + -203, + 170, + 206, + 228, + 216, + 160, + 218, + -203, + -203, + 19, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + 207, + 77, + 25, + 140, + 140, + 110, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + 33, + 111, + 222, + 168, + -203, + 226, + 120, + 247, + 229, + 109, + 109, + 246, + 243, + 204, + 59, + -203, + 232, + -203, + -203, + -203, + -203, + 140, + 202, + 202, + -13, + -13, + -203, + 179, + 214, + -203, + -203, + -203, + 178, + 199, + -203, + 120, + -203, + 180, + 120, + 203, + 134, + 7, + -203, + 140, + 202, + 249, + 181, + -203, + 160, + 160, + -13, + -203, + 208, + 238, + -203, + -203, + 20, + 239, + -203, + 240, + 202, + 228, + -203, + 25, + 260, + 223, + 218, + -203, + 105, + 54, + 180, + -203, + 220, + -203, + -23, + -203, + 140, + 189, + -203, + -203, + -203, + -203, + 245, + 209, + 22, + -203, + 242, + -203, + 101, + -203, + 109, + -203, + -203, + 140, + 200, + 219, + -203, + -203, + 210, + 224, + 264, + -203, + 265, + 241, + 230, + 234, + 235, + 224, + 233, + 107, + 267, + -203, + 248, + 250, + 236, + 244, + 160, + 160, + 285, + 289, + 251, + 252, + 253, + 254, + 160, + 160, + 284, + 294, + -203, + -203}; /* YYDEFACT[STATE-NUM] -- Default reduction number in state STATE-NUM. Performed when YYTABLE does not specify something else to do. Zero means the default is an error. */ -static const yytype_int8 yydefact[] = -{ - 0, 0, 0, 0, 0, 0, 0, 25, 0, 0, - 0, 26, 27, 28, 24, 23, 0, 0, 0, 0, - 91, 22, 21, 14, 15, 16, 17, 9, 10, 11, - 12, 13, 8, 5, 7, 6, 4, 3, 18, 19, - 20, 0, 0, 0, 0, 0, 47, 48, 67, 49, - 0, 66, 64, 55, 56, 65, 0, 31, 30, 0, - 0, 0, 0, 0, 89, 1, 92, 2, 0, 0, - 29, 0, 0, 0, 63, 0, 0, 0, 0, 0, - 0, 0, 72, 0, 0, 0, 0, 0, 0, 62, - 68, 57, 58, 59, 60, 61, 69, 70, 72, 0, - 74, 52, 0, 90, 0, 0, 35, 0, 33, 0, - 87, 0, 0, 0, 73, 75, 0, 0, 40, 41, - 42, 43, 38, 0, 0, 0, 71, 54, 45, 81, - 82, 83, 84, 85, 86, 0, 0, 74, 72, 0, - 0, 35, 50, 0, 0, 0, 78, 80, 77, 79, - 76, 53, 88, 39, 0, 36, 0, 34, 32, 45, - 44, 37, 0, 46, 0, 51 -}; +static const yytype_uint8 yydefact[] = {0, + 38, + 0, + 103, + 103, + 0, + 0, + 27, + 0, + 0, + 0, + 28, + 29, + 30, + 26, + 25, + 0, + 0, + 0, + 0, + 24, + 23, + 17, + 18, + 19, + 20, + 9, + 10, + 11, + 14, + 12, + 13, + 8, + 15, + 16, + 5, + 7, + 6, + 3, + 4, + 21, + 22, + 0, + 0, + 37, + 0, + 0, + 0, + 0, + 0, + 103, + 76, + 88, + 89, + 90, + 85, + 86, + 123, + 87, + 0, + 114, + 112, + 101, + 118, + 116, + 117, + 113, + 102, + 33, + 32, + 0, + 0, + 0, + 0, + 0, + 163, + 1, + 165, + 2, + 92, + 0, + 0, + 0, + 31, + 0, + 53, + 103, + 0, + 0, + 72, + 74, + 0, + 77, + 0, + 103, + 0, + 111, + 0, + 119, + 0, + 0, + 0, + 0, + 104, + 0, + 0, + 0, + 130, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 122, + 110, + 73, + 75, + 91, + 0, + 0, + 124, + 115, + 120, + 106, + 107, + 108, + 109, + 103, + 125, + 118, + 130, + 34, + 0, + 0, + 0, + 94, + 0, + 130, + 96, + 0, + 164, + 82, + 0, + 54, + 0, + 0, + 50, + 0, + 51, + 43, + 0, + 0, + 45, + 78, + 121, + 105, + 0, + 126, + 157, + 0, + 79, + 68, + 148, + 146, + 0, + 136, + 137, + 138, + 139, + 140, + 141, + 144, + 142, + 0, + 131, + 0, + 0, + 0, + 95, + 83, + 84, + 62, + 63, + 64, + 65, + 66, + 67, + 61, + 0, + 0, + 0, + 49, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 159, + 0, + 0, + 80, + 0, + 149, + 147, + 145, + 143, + 0, + 0, + 0, + 133, + 98, + 97, + 0, + 0, + 60, + 59, + 57, + 54, + 92, + 93, + 0, + 44, + 0, + 0, + 0, + 130, + 118, + 127, + 103, + 0, + 150, + 0, + 70, + 0, + 79, + 132, + 134, + 135, + 0, + 58, + 55, + 48, + 0, + 52, + 0, + 0, + 157, + 158, + 160, + 0, + 161, + 69, + 81, + 0, + 61, + 0, + 47, + 0, + 35, + 128, + 100, + 0, + 0, + 99, + 71, + 56, + 46, + 0, + 0, + 154, + 151, + 152, + 162, + 0, + 36, + 0, + 156, + 155, + 0, + 0, + 0, + 129, + 153, + 0, + 0, + 0, + 39, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 40, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 0, + 41, + 42}; /* YYPGOTO[NTERM-NUM]. */ -static const yytype_int8 yypgoto[] = -{ - -97, -97, 115, -97, -97, -97, -97, -97, -97, -97, - -97, -97, -97, -97, -97, -1, 11, -97, -97, -97, - -22, -83, -97, -97, -97, -97, -97, -4, 27, -77, - -97, 26, -96, 1, -97, 28, -97, -97, -97, -97, - -97 -}; +static const yytype_int16 yypgoto[] = {-203, + -203, + 305, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + -203, + 23, + -203, + -130, + -203, + -203, + -203, + -203, + 103, + 132, + 68, + -203, + -203, + 92, + 211, + -203, + 94, + -100, + -102, + 112, + -203, + -203, + -203, + 149, + -49, + -203, + -4, + -57, + 270, + -203, + -203, + -203, + -99, + 131, + 56, + -133, + -202, + 162, + -203, + 60, + -203, + 89, + -203, + -203, + -203, + -203, + -203}; /* YYDEFGOTO[NTERM-NUM]. */ -static const yytype_uint8 yydefgoto[] = -{ - 0, 19, 20, 21, 22, 23, 24, 25, 26, 27, - 28, 29, 30, 31, 32, 124, 106, 154, 122, 33, - 145, 52, 157, 34, 35, 36, 37, 53, 54, 55, - 97, 98, 101, 114, 115, 135, 127, 38, 39, 40, - 67 -}; +static const yytype_int16 yydefgoto[] = {0, + 18, + 19, + 20, + 21, + 22, + 23, + 24, + 25, + 26, + 27, + 28, + 29, + 30, + 46, + 295, + 282, + 156, + 31, + 32, + 33, + 34, + 195, + 149, + 224, + 193, + 35, + 167, + 92, + 93, + 207, + 208, + 61, + 112, + 36, + 37, + 143, + 144, + 38, + 39, + 62, + 63, + 163, + 64, + 65, + 66, + 232, + 136, + 233, + 141, + 180, + 181, + 258, + 278, + 279, + 205, + 238, + 271, + 40, + 41, + 78}; /* YYTABLE[YYPACT[STATE-NUM]] -- What to do in state STATE-NUM. If positive, shift that token. If negative, reduce the rule whose number is the opposite. If YYTABLE_NINF, syntax error. */ -static const yytype_uint8 yytable[] = -{ - 56, 103, 110, 45, 46, 47, 48, 49, 89, 46, - 47, 58, 49, 75, 41, 57, 42, 112, 118, 119, - 120, 121, 43, 113, 44, 78, 79, 59, 128, 61, - 62, 60, 65, 138, 63, 46, 47, 48, 49, 66, - 50, 51, 151, 76, 77, 78, 79, 76, 77, 78, - 79, 68, 146, 148, 112, 73, 1, 2, 147, 149, - 113, 159, 3, 4, 5, 6, 7, 8, 9, 10, - 69, 91, 72, 11, 12, 13, 70, 74, 71, 80, - 14, 15, 129, 130, 131, 132, 133, 134, 16, 83, - 17, 81, 82, 18, 84, 85, 86, 87, 88, 90, - 96, 99, 100, 92, 93, 94, 95, 104, 102, 105, - 107, 108, 109, 111, 116, 117, 123, 125, 137, 139, - 140, 144, 142, 143, 152, 153, 158, 156, 160, 161, - 162, 165, 164, 64, 141, 126, 0, 163, 150, 0, - 155, 136 -}; - -static const yytype_int16 yycheck[] = -{ - 4, 84, 98, 19, 51, 52, 53, 54, 20, 51, - 52, 9, 54, 21, 8, 53, 10, 100, 25, 26, - 27, 28, 8, 100, 10, 57, 58, 32, 111, 53, - 53, 34, 0, 116, 40, 51, 52, 53, 54, 3, - 56, 57, 138, 55, 56, 57, 58, 55, 56, 57, - 58, 53, 135, 136, 137, 31, 5, 6, 135, 136, - 137, 144, 11, 12, 13, 14, 15, 16, 17, 18, - 53, 75, 45, 22, 23, 24, 53, 50, 53, 34, - 29, 30, 45, 46, 47, 48, 49, 50, 37, 37, - 39, 53, 53, 42, 45, 41, 19, 38, 38, 53, - 53, 33, 35, 76, 77, 78, 79, 54, 53, 53, - 53, 53, 21, 19, 45, 32, 21, 19, 36, 8, - 19, 21, 20, 53, 53, 51, 20, 43, 20, 20, - 44, 53, 45, 18, 123, 109, -1, 159, 137, -1, - 141, 113 -}; +static const yytype_int16 yytable[] = {67, + 87, + 96, + 164, + 97, + 135, + 137, + 147, + 165, + 146, + 184, + 97, + 97, + 244, + 245, + 68, + 94, + 47, + 110, + 48, + 151, + 138, + 123, + 124, + 263, + 126, + 42, + 284, + 210, + 127, + 114, + 215, + 216, + 86, + 183, + 256, + 71, + 86, + 73, + 276, + 285, + 95, + 129, + 130, + 131, + 132, + 88, + 74, + 139, + 86, + 115, + 267, + 211, + 43, + 111, + 44, + 50, + 220, + 51, + 140, + 69, + 70, + 162, + 152, + 147, + 154, + 52, + 53, + 49, + 229, + 99, + 100, + 101, + 102, + 221, + 54, + 222, + 45, + 223, + 215, + 216, + 98, + 118, + 179, + 240, + 76, + 168, + 241, + 98, + 98, + 125, + 99, + 100, + 101, + 102, + 221, + 212, + 222, + 250, + 223, + 254, + 252, + 197, + 72, + 234, + 99, + 100, + 101, + 102, + 77, + 169, + 55, + 56, + 57, + 58, + 187, + 59, + 60, + 170, + 188, + 189, + 190, + 191, + 192, + 217, + 218, + 89, + 90, + 80, + 161, + 272, + 91, + 304, + 241, + 79, + 305, + 121, + 122, + 82, + 147, + 147, + 260, + 81, + 101, + 102, + 185, + 186, + 104, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 83, + 243, + 179, + 179, + 99, + 100, + 101, + 102, + 50, + 84, + 51, + 287, + 288, + 85, + 105, + 106, + 107, + 108, + 52, + 53, + 117, + 113, + 116, + 109, + 179, + 251, + 119, + 54, + 120, + 128, + 51, + 140, + 150, + 133, + 134, + 142, + 148, + 153, + 52, + 53, + 179, + 86, + 157, + 160, + 264, + 155, + 158, + 54, + 182, + 166, + 194, + 147, + 147, + 311, + 312, + 168, + 277, + 196, + 274, + 147, + 147, + 319, + 320, + 55, + 56, + 57, + 58, + 199, + 59, + 60, + 50, + 198, + 51, + 277, + 200, + 201, + 255, + 202, + 203, + 169, + 52, + 53, + 204, + 55, + 56, + 206, + 58, + 170, + 145, + 54, + 209, + 226, + 213, + 227, + 228, + 230, + 236, + 231, + 237, + 239, + 242, + 247, + 246, + 111, + 253, + 257, + 215, + 262, + 265, + 266, + 269, + 270, + 280, + 281, + 286, + 291, + 283, + 171, + 172, + 173, + 174, + 175, + 176, + 177, + 178, + 55, + 56, + 57, + 58, + 275, + 59, + 60, + 1, + 2, + 292, + 293, + 296, + 297, + 294, + 306, + 299, + 3, + 4, + 5, + 6, + 7, + 8, + 9, + 10, + 300, + 301, + 309, + 298, + 321, + 11, + 12, + 13, + 313, + 303, + 310, + 307, + 314, + 308, + 322, + 315, + 316, + 75, + 302, + 317, + 318, + 225, + 14, + 248, + 15, + 273, + 259, + 219, + 103, + 235, + 159, + 261, + 16, + 249, + 289, + 17, + 214, + 0, + 268, + 0, + 0, + 290}; + +static const yytype_int16 yycheck[] = {4, + 50, + 59, + 136, + 4, + 104, + 105, + 109, + 138, + 109, + 143, + 4, + 4, + 215, + 216, + 81, + 24, + 13, + 24, + 15, + 4, + 24, + 27, + 28, + 4, + 81, + 13, + 5, + 9, + 85, + 4, + 54, + 55, + 17, + 28, + 237, + 50, + 17, + 81, + 62, + 18, + 49, + 99, + 100, + 101, + 102, + 50, + 81, + 51, + 17, + 24, + 253, + 33, + 40, + 60, + 42, + 24, + 24, + 26, + 53, + 14, + 15, + 62, + 112, + 166, + 114, + 34, + 35, + 64, + 199, + 83, + 84, + 85, + 86, + 41, + 43, + 43, + 64, + 45, + 54, + 55, + 81, + 86, + 140, + 25, + 0, + 9, + 28, + 81, + 81, + 94, + 83, + 84, + 85, + 86, + 41, + 77, + 43, + 228, + 45, + 233, + 231, + 151, + 52, + 203, + 83, + 84, + 85, + 86, + 3, + 33, + 79, + 80, + 81, + 82, + 32, + 84, + 85, + 41, + 36, + 37, + 38, + 39, + 40, + 181, + 182, + 79, + 80, + 15, + 133, + 25, + 84, + 25, + 28, + 81, + 28, + 79, + 80, + 15, + 241, + 242, + 241, + 81, + 85, + 86, + 79, + 80, + 52, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 81, + 214, + 215, + 216, + 83, + 84, + 85, + 86, + 24, + 81, + 26, + 66, + 67, + 81, + 52, + 81, + 81, + 56, + 34, + 35, + 57, + 81, + 81, + 71, + 237, + 230, + 25, + 43, + 25, + 81, + 26, + 53, + 61, + 28, + 81, + 81, + 81, + 57, + 34, + 35, + 253, + 17, + 57, + 25, + 249, + 81, + 81, + 43, + 71, + 24, + 28, + 309, + 310, + 309, + 310, + 9, + 269, + 71, + 263, + 317, + 318, + 317, + 318, + 79, + 80, + 81, + 82, + 28, + 84, + 85, + 24, + 81, + 26, + 286, + 25, + 81, + 236, + 63, + 28, + 33, + 34, + 35, + 10, + 79, + 80, + 25, + 82, + 41, + 84, + 43, + 28, + 25, + 41, + 81, + 24, + 4, + 6, + 24, + 11, + 51, + 24, + 43, + 79, + 60, + 57, + 12, + 54, + 25, + 25, + 25, + 6, + 44, + 79, + 24, + 28, + 71, + 63, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 79, + 80, + 81, + 82, + 65, + 84, + 85, + 7, + 8, + 71, + 81, + 28, + 28, + 70, + 28, + 66, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 71, + 71, + 71, + 67, + 25, + 29, + 30, + 31, + 28, + 81, + 71, + 68, + 28, + 68, + 25, + 69, + 69, + 17, + 300, + 71, + 71, + 194, + 46, + 225, + 48, + 262, + 239, + 183, + 63, + 203, + 124, + 242, + 56, + 226, + 283, + 59, + 179, + -1, + 254, + -1, + -1, + 286}; /* YYSTOS[STATE-NUM] -- The symbol kind of the accessing symbol of state STATE-NUM. */ -static const yytype_int8 yystos[] = -{ - 0, 5, 6, 11, 12, 13, 14, 15, 16, 17, - 18, 22, 23, 24, 29, 30, 37, 39, 42, 61, - 62, 63, 64, 65, 66, 67, 68, 69, 70, 71, - 72, 73, 74, 79, 83, 84, 85, 86, 97, 98, - 99, 8, 10, 8, 10, 19, 51, 52, 53, 54, - 56, 57, 81, 87, 88, 89, 87, 53, 9, 32, - 34, 53, 53, 40, 62, 0, 3, 100, 53, 53, - 53, 53, 88, 31, 88, 21, 55, 56, 57, 58, - 34, 53, 53, 37, 45, 41, 19, 38, 38, 20, - 53, 87, 88, 88, 88, 88, 53, 90, 91, 33, - 35, 92, 53, 81, 54, 53, 76, 53, 53, 21, - 92, 19, 81, 89, 93, 94, 45, 32, 25, 26, - 27, 28, 78, 21, 75, 19, 91, 96, 81, 45, - 46, 47, 48, 49, 50, 95, 95, 36, 81, 8, - 19, 76, 20, 53, 21, 80, 81, 89, 81, 89, - 93, 92, 53, 51, 77, 75, 43, 82, 20, 81, - 20, 20, 44, 80, 45, 53 -}; +static const yytype_uint8 yystos[] = {0, + 7, + 8, + 16, + 17, + 18, + 19, + 20, + 21, + 22, + 23, + 29, + 30, + 31, + 46, + 48, + 56, + 59, + 89, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 106, + 107, + 108, + 109, + 114, + 122, + 123, + 126, + 127, + 146, + 147, + 13, + 40, + 42, + 64, + 102, + 13, + 15, + 64, + 24, + 26, + 34, + 35, + 43, + 79, + 80, + 81, + 82, + 84, + 85, + 120, + 128, + 129, + 131, + 132, + 133, + 128, + 81, + 14, + 15, + 50, + 52, + 81, + 81, + 90, + 0, + 3, + 148, + 81, + 15, + 81, + 15, + 81, + 81, + 81, + 17, + 126, + 128, + 79, + 80, + 84, + 116, + 117, + 24, + 49, + 129, + 4, + 81, + 83, + 84, + 85, + 86, + 130, + 52, + 52, + 81, + 81, + 56, + 71, + 24, + 60, + 121, + 81, + 4, + 24, + 81, + 57, + 128, + 25, + 25, + 79, + 80, + 27, + 28, + 128, + 81, + 85, + 81, + 129, + 129, + 129, + 129, + 28, + 81, + 134, + 135, + 134, + 24, + 51, + 53, + 137, + 81, + 124, + 125, + 84, + 119, + 120, + 81, + 111, + 61, + 4, + 126, + 57, + 126, + 81, + 105, + 57, + 81, + 116, + 25, + 128, + 62, + 130, + 137, + 105, + 24, + 115, + 9, + 33, + 41, + 71, + 72, + 73, + 74, + 75, + 76, + 77, + 78, + 129, + 138, + 139, + 71, + 28, + 137, + 79, + 80, + 32, + 36, + 37, + 38, + 39, + 40, + 113, + 28, + 110, + 71, + 126, + 81, + 28, + 25, + 81, + 63, + 28, + 10, + 143, + 25, + 118, + 119, + 28, + 9, + 33, + 77, + 41, + 139, + 54, + 55, + 129, + 129, + 125, + 24, + 41, + 43, + 45, + 112, + 111, + 25, + 81, + 24, + 105, + 4, + 24, + 134, + 136, + 134, + 135, + 6, + 11, + 144, + 51, + 25, + 28, + 24, + 129, + 138, + 138, + 79, + 43, + 110, + 121, + 105, + 126, + 105, + 57, + 137, + 128, + 138, + 12, + 140, + 115, + 119, + 118, + 25, + 4, + 126, + 25, + 25, + 138, + 143, + 6, + 44, + 145, + 25, + 112, + 126, + 65, + 62, + 129, + 141, + 142, + 79, + 24, + 104, + 63, + 5, + 18, + 28, + 66, + 67, + 136, + 141, + 71, + 71, + 81, + 70, + 103, + 28, + 28, + 67, + 66, + 71, + 71, + 103, + 81, + 25, + 28, + 28, + 68, + 68, + 71, + 71, + 119, + 119, + 28, + 28, + 69, + 69, + 71, + 71, + 119, + 119, + 25, + 25}; /* YYR1[RULE-NUM] -- Symbol kind of the left-hand side of rule RULE-NUM. */ -static const yytype_int8 yyr1[] = -{ - 0, 60, 61, 62, 62, 62, 62, 62, 62, 62, - 62, 62, 62, 62, 62, 62, 62, 62, 62, 62, - 62, 62, 62, 63, 64, 65, 66, 67, 68, 69, - 70, 71, 72, 73, 74, 75, 75, 76, 76, 77, - 78, 78, 78, 78, 79, 80, 80, 81, 81, 81, - 82, 82, 83, 84, 85, 86, 87, 87, 88, 88, - 88, 88, 88, 88, 88, 88, 88, 89, 89, 90, - 91, 91, 92, 92, 93, 93, 93, 94, 94, 94, - 94, 95, 95, 95, 95, 95, 95, 96, 97, 98, - 99, 100, 100 -}; +static const yytype_uint8 yyr1[] = {0, + 88, + 89, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 90, + 91, + 92, + 93, + 94, + 95, + 96, + 97, + 98, + 99, + 100, + 101, + 101, + 102, + 102, + 103, + 104, + 104, + 104, + 105, + 105, + 106, + 107, + 107, + 107, + 107, + 107, + 108, + 108, + 109, + 110, + 110, + 111, + 111, + 112, + 112, + 112, + 112, + 113, + 113, + 113, + 113, + 113, + 113, + 114, + 114, + 115, + 115, + 116, + 116, + 116, + 116, + 117, + 117, + 117, + 118, + 118, + 118, + 119, + 119, + 119, + 120, + 120, + 120, + 120, + 120, + 120, + 120, + 121, + 121, + 122, + 123, + 124, + 124, + 125, + 126, + 126, + 127, + 127, + 128, + 128, + 128, + 129, + 129, + 129, + 129, + 129, + 129, + 129, + 129, + 129, + 129, + 129, + 129, + 130, + 130, + 130, + 131, + 132, + 133, + 133, + 134, + 135, + 135, + 136, + 136, + 137, + 137, + 138, + 138, + 138, + 138, + 139, + 139, + 139, + 139, + 139, + 139, + 139, + 139, + 139, + 139, + 139, + 139, + 139, + 139, + 140, + 140, + 141, + 141, + 142, + 142, + 142, + 143, + 143, + 144, + 144, + 145, + 145, + 146, + 147, + 148}; /* YYR2[RULE-NUM] -- Number of symbols on the right-hand side of rule RULE-NUM. */ -static const yytype_int8 yyr2[] = +static const yytype_int8 yyr2[] = {0, + 2, + 2, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 3, + 2, + 2, + 4, + 9, + 11, + 1, + 0, + 1, + 9, + 17, + 17, + 1, + 3, + 5, + 10, + 9, + 8, + 6, + 5, + 5, + 8, + 3, + 0, + 3, + 6, + 3, + 2, + 1, + 1, + 0, + 1, + 1, + 1, + 1, + 1, + 1, + 5, + 8, + 3, + 5, + 1, + 2, + 1, + 2, + 0, + 1, + 3, + 0, + 1, + 3, + 1, + 2, + 2, + 1, + 1, + 1, + 1, + 1, + 1, + 3, + 0, + 4, + 4, + 5, + 1, + 3, + 3, + 9, + 9, + 2, + 2, + 0, + 2, + 4, + 3, + 3, + 3, + 3, + 3, + 2, + 1, + 1, + 1, + 3, + 1, + 1, + 0, + 1, + 2, + 4, + 3, + 1, + 3, + 1, + 2, + 4, + 3, + 6, + 0, + 2, + 3, + 2, + 3, + 3, + 1, + 1, + 1, + 1, + 1, + 1, + 1, + 2, + 1, + 2, + 1, + 2, + 1, + 2, + 0, + 3, + 1, + 3, + 1, + 2, + 2, + 0, + 3, + 0, + 2, + 0, + 2, + 2, + 4, + 1}; + +enum { - 0, 2, 2, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, - 1, 1, 1, 1, 1, 1, 1, 1, 1, 3, - 2, 2, 8, 5, 8, 0, 3, 5, 2, 1, - 1, 1, 1, 1, 8, 0, 3, 1, 1, 1, - 0, 4, 4, 7, 6, 2, 1, 3, 3, 3, - 3, 3, 3, 2, 1, 1, 1, 1, 3, 1, - 1, 3, 0, 2, 0, 1, 3, 3, 3, 3, - 3, 1, 1, 1, 1, 1, 1, 0, 7, 2, - 4, 0, 1 + YYENOMEM = -2 }; - -enum { YYENOMEM = -2 }; - -#define yyerrok (yyerrstatus = 0) -#define yyclearin (yychar = YYEMPTY) - -#define YYACCEPT goto yyacceptlab -#define YYABORT goto yyabortlab -#define YYERROR goto yyerrorlab -#define YYNOMEM goto yyexhaustedlab - - -#define YYRECOVERING() (!!yyerrstatus) - -#define YYBACKUP(Token, Value) \ - do \ - if (yychar == YYEMPTY) \ - { \ - yychar = (Token); \ - yylval = (Value); \ - YYPOPSTACK (yylen); \ - yystate = *yyssp; \ - goto yybackup; \ - } \ - else \ - { \ - yyerror (&yylloc, sql_string, sql_result, scanner, YY_("syntax error: cannot back up")); \ - YYERROR; \ - } \ +#define yyerrok (yyerrstatus = 0) +#define yyclearin (yychar = YYEMPTY) + +#define YYACCEPT goto yyacceptlab +#define YYABORT goto yyabortlab +#define YYERROR goto yyerrorlab +#define YYNOMEM goto yyexhaustedlab + +#define YYRECOVERING() (!!yyerrstatus) + +#define YYBACKUP(Token, Value) \ + do \ + if (yychar == YYEMPTY) { \ + yychar = (Token); \ + yylval = (Value); \ + YYPOPSTACK(yylen); \ + yystate = *yyssp; \ + goto yybackup; \ + } else { \ + yyerror(&yylloc, sql_string, sql_result, scanner, YY_("syntax error: cannot back up")); \ + YYERROR; \ + } \ while (0) /* Backward compatibility with an undocumented macro. @@ -906,151 +3503,131 @@ enum { YYENOMEM = -2 }; the previous symbol: RHS[0] (always defined). */ #ifndef YYLLOC_DEFAULT -# define YYLLOC_DEFAULT(Current, Rhs, N) \ - do \ - if (N) \ - { \ - (Current).first_line = YYRHSLOC (Rhs, 1).first_line; \ - (Current).first_column = YYRHSLOC (Rhs, 1).first_column; \ - (Current).last_line = YYRHSLOC (Rhs, N).last_line; \ - (Current).last_column = YYRHSLOC (Rhs, N).last_column; \ - } \ - else \ - { \ - (Current).first_line = (Current).last_line = \ - YYRHSLOC (Rhs, 0).last_line; \ - (Current).first_column = (Current).last_column = \ - YYRHSLOC (Rhs, 0).last_column; \ - } \ - while (0) +#define YYLLOC_DEFAULT(Current, Rhs, N) \ + do \ + if (N) { \ + (Current).first_line = YYRHSLOC(Rhs, 1).first_line; \ + (Current).first_column = YYRHSLOC(Rhs, 1).first_column; \ + (Current).last_line = YYRHSLOC(Rhs, N).last_line; \ + (Current).last_column = YYRHSLOC(Rhs, N).last_column; \ + } else { \ + (Current).first_line = (Current).last_line = YYRHSLOC(Rhs, 0).last_line; \ + (Current).first_column = (Current).last_column = YYRHSLOC(Rhs, 0).last_column; \ + } \ + while (0) #endif #define YYRHSLOC(Rhs, K) ((Rhs)[K]) - /* Enable debugging if requested. */ #if YYDEBUG -# ifndef YYFPRINTF -# include /* INFRINGES ON USER NAME SPACE */ -# define YYFPRINTF fprintf -# endif - -# define YYDPRINTF(Args) \ -do { \ - if (yydebug) \ - YYFPRINTF Args; \ -} while (0) +#ifndef YYFPRINTF +#include /* INFRINGES ON USER NAME SPACE */ +#define YYFPRINTF fprintf +#endif +#define YYDPRINTF(Args) \ + do { \ + if (yydebug) \ + YYFPRINTF Args; \ + } while (0) /* YYLOCATION_PRINT -- Print the location on the stream. This macro was not mandated originally: define only if we know we won't break user code: when these are the locations we know. */ -# ifndef YYLOCATION_PRINT +#ifndef YYLOCATION_PRINT -# if defined YY_LOCATION_PRINT +#if defined YY_LOCATION_PRINT - /* Temporary convenience wrapper in case some people defined the - undocumented and private YY_LOCATION_PRINT macros. */ -# define YYLOCATION_PRINT(File, Loc) YY_LOCATION_PRINT(File, *(Loc)) +/* Temporary convenience wrapper in case some people defined the + undocumented and private YY_LOCATION_PRINT macros. */ +#define YYLOCATION_PRINT(File, Loc) YY_LOCATION_PRINT(File, *(Loc)) -# elif defined YYLTYPE_IS_TRIVIAL && YYLTYPE_IS_TRIVIAL +#elif defined YYLTYPE_IS_TRIVIAL && YYLTYPE_IS_TRIVIAL /* Print *YYLOCP on YYO. Private, do not rely on its existence. */ YY_ATTRIBUTE_UNUSED -static int -yy_location_print_ (FILE *yyo, YYLTYPE const * const yylocp) +static int yy_location_print_(FILE *yyo, YYLTYPE const *const yylocp) { - int res = 0; + int res = 0; int end_col = 0 != yylocp->last_column ? yylocp->last_column - 1 : 0; - if (0 <= yylocp->first_line) - { - res += YYFPRINTF (yyo, "%d", yylocp->first_line); - if (0 <= yylocp->first_column) - res += YYFPRINTF (yyo, ".%d", yylocp->first_column); - } - if (0 <= yylocp->last_line) - { - if (yylocp->first_line < yylocp->last_line) - { - res += YYFPRINTF (yyo, "-%d", yylocp->last_line); - if (0 <= end_col) - res += YYFPRINTF (yyo, ".%d", end_col); - } - else if (0 <= end_col && yylocp->first_column < end_col) - res += YYFPRINTF (yyo, "-%d", end_col); - } + if (0 <= yylocp->first_line) { + res += YYFPRINTF(yyo, "%d", yylocp->first_line); + if (0 <= yylocp->first_column) + res += YYFPRINTF(yyo, ".%d", yylocp->first_column); + } + if (0 <= yylocp->last_line) { + if (yylocp->first_line < yylocp->last_line) { + res += YYFPRINTF(yyo, "-%d", yylocp->last_line); + if (0 <= end_col) + res += YYFPRINTF(yyo, ".%d", end_col); + } else if (0 <= end_col && yylocp->first_column < end_col) + res += YYFPRINTF(yyo, "-%d", end_col); + } return res; } -# define YYLOCATION_PRINT yy_location_print_ +#define YYLOCATION_PRINT yy_location_print_ - /* Temporary convenience wrapper in case some people defined the - undocumented and private YY_LOCATION_PRINT macros. */ -# define YY_LOCATION_PRINT(File, Loc) YYLOCATION_PRINT(File, &(Loc)) +/* Temporary convenience wrapper in case some people defined the + undocumented and private YY_LOCATION_PRINT macros. */ +#define YY_LOCATION_PRINT(File, Loc) YYLOCATION_PRINT(File, &(Loc)) -# else - -# define YYLOCATION_PRINT(File, Loc) ((void) 0) - /* Temporary convenience wrapper in case some people defined the - undocumented and private YY_LOCATION_PRINT macros. */ -# define YY_LOCATION_PRINT YYLOCATION_PRINT - -# endif -# endif /* !defined YYLOCATION_PRINT */ +#else +#define YYLOCATION_PRINT(File, Loc) ((void)0) +/* Temporary convenience wrapper in case some people defined the + undocumented and private YY_LOCATION_PRINT macros. */ +#define YY_LOCATION_PRINT YYLOCATION_PRINT -# define YY_SYMBOL_PRINT(Title, Kind, Value, Location) \ -do { \ - if (yydebug) \ - { \ - YYFPRINTF (stderr, "%s ", Title); \ - yy_symbol_print (stderr, \ - Kind, Value, Location, sql_string, sql_result, scanner); \ - YYFPRINTF (stderr, "\n"); \ - } \ -} while (0) +#endif +#endif /* !defined YYLOCATION_PRINT */ +#define YY_SYMBOL_PRINT(Title, Kind, Value, Location) \ + do { \ + if (yydebug) { \ + YYFPRINTF(stderr, "%s ", Title); \ + yy_symbol_print(stderr, Kind, Value, Location, sql_string, sql_result, scanner); \ + YYFPRINTF(stderr, "\n"); \ + } \ + } while (0) /*-----------------------------------. | Print this symbol's value on YYO. | `-----------------------------------*/ -static void -yy_symbol_value_print (FILE *yyo, - yysymbol_kind_t yykind, YYSTYPE const * const yyvaluep, YYLTYPE const * const yylocationp, const char * sql_string, ParsedSqlResult * sql_result, void * scanner) +static void yy_symbol_value_print(FILE *yyo, yysymbol_kind_t yykind, YYSTYPE const *const yyvaluep, + YYLTYPE const *const yylocationp, const char *sql_string, ParsedSqlResult *sql_result, void *scanner) { FILE *yyoutput = yyo; - YY_USE (yyoutput); - YY_USE (yylocationp); - YY_USE (sql_string); - YY_USE (sql_result); - YY_USE (scanner); + YY_USE(yyoutput); + YY_USE(yylocationp); + YY_USE(sql_string); + YY_USE(sql_result); + YY_USE(scanner); if (!yyvaluep) return; YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN - YY_USE (yykind); + YY_USE(yykind); YY_IGNORE_MAYBE_UNINITIALIZED_END } - /*---------------------------. | Print this symbol on YYO. | `---------------------------*/ -static void -yy_symbol_print (FILE *yyo, - yysymbol_kind_t yykind, YYSTYPE const * const yyvaluep, YYLTYPE const * const yylocationp, const char * sql_string, ParsedSqlResult * sql_result, void * scanner) +static void yy_symbol_print(FILE *yyo, yysymbol_kind_t yykind, YYSTYPE const *const yyvaluep, + YYLTYPE const *const yylocationp, const char *sql_string, ParsedSqlResult *sql_result, void *scanner) { - YYFPRINTF (yyo, "%s %s (", - yykind < YYNTOKENS ? "token" : "nterm", yysymbol_name (yykind)); + YYFPRINTF(yyo, "%s %s (", yykind < YYNTOKENS ? "token" : "nterm", yysymbol_name(yykind)); - YYLOCATION_PRINT (yyo, yylocationp); - YYFPRINTF (yyo, ": "); - yy_symbol_value_print (yyo, yykind, yyvaluep, yylocationp, sql_string, sql_result, scanner); - YYFPRINTF (yyo, ")"); + YYLOCATION_PRINT(yyo, yylocationp); + YYFPRINTF(yyo, ": "); + yy_symbol_value_print(yyo, yykind, yyvaluep, yylocationp, sql_string, sql_result, scanner); + YYFPRINTF(yyo, ")"); } /*------------------------------------------------------------------. @@ -1058,70 +3635,66 @@ yy_symbol_print (FILE *yyo, | TOP (included). | `------------------------------------------------------------------*/ -static void -yy_stack_print (yy_state_t *yybottom, yy_state_t *yytop) +static void yy_stack_print(yy_state_t *yybottom, yy_state_t *yytop) { - YYFPRINTF (stderr, "Stack now"); - for (; yybottom <= yytop; yybottom++) - { - int yybot = *yybottom; - YYFPRINTF (stderr, " %d", yybot); - } - YYFPRINTF (stderr, "\n"); + YYFPRINTF(stderr, "Stack now"); + for (; yybottom <= yytop; yybottom++) { + int yybot = *yybottom; + YYFPRINTF(stderr, " %d", yybot); + } + YYFPRINTF(stderr, "\n"); } -# define YY_STACK_PRINT(Bottom, Top) \ -do { \ - if (yydebug) \ - yy_stack_print ((Bottom), (Top)); \ -} while (0) - +#define YY_STACK_PRINT(Bottom, Top) \ + do { \ + if (yydebug) \ + yy_stack_print((Bottom), (Top)); \ + } while (0) /*------------------------------------------------. | Report that the YYRULE is going to be reduced. | `------------------------------------------------*/ -static void -yy_reduce_print (yy_state_t *yyssp, YYSTYPE *yyvsp, YYLTYPE *yylsp, - int yyrule, const char * sql_string, ParsedSqlResult * sql_result, void * scanner) +static void yy_reduce_print(yy_state_t *yyssp, YYSTYPE *yyvsp, YYLTYPE *yylsp, int yyrule, const char *sql_string, + ParsedSqlResult *sql_result, void *scanner) { - int yylno = yyrline[yyrule]; + int yylno = yyrline[yyrule]; int yynrhs = yyr2[yyrule]; int yyi; - YYFPRINTF (stderr, "Reducing stack by rule %d (line %d):\n", - yyrule - 1, yylno); + YYFPRINTF(stderr, "Reducing stack by rule %d (line %d):\n", yyrule - 1, yylno); /* The symbols being reduced. */ - for (yyi = 0; yyi < yynrhs; yyi++) - { - YYFPRINTF (stderr, " $%d = ", yyi + 1); - yy_symbol_print (stderr, - YY_ACCESSING_SYMBOL (+yyssp[yyi + 1 - yynrhs]), - &yyvsp[(yyi + 1) - (yynrhs)], - &(yylsp[(yyi + 1) - (yynrhs)]), sql_string, sql_result, scanner); - YYFPRINTF (stderr, "\n"); - } + for (yyi = 0; yyi < yynrhs; yyi++) { + YYFPRINTF(stderr, " $%d = ", yyi + 1); + yy_symbol_print(stderr, + YY_ACCESSING_SYMBOL(+yyssp[yyi + 1 - yynrhs]), + &yyvsp[(yyi + 1) - (yynrhs)], + &(yylsp[(yyi + 1) - (yynrhs)]), + sql_string, + sql_result, + scanner); + YYFPRINTF(stderr, "\n"); + } } -# define YY_REDUCE_PRINT(Rule) \ -do { \ - if (yydebug) \ - yy_reduce_print (yyssp, yyvsp, yylsp, Rule, sql_string, sql_result, scanner); \ -} while (0) +#define YY_REDUCE_PRINT(Rule) \ + do { \ + if (yydebug) \ + yy_reduce_print(yyssp, yyvsp, yylsp, Rule, sql_string, sql_result, scanner); \ + } while (0) /* Nonzero means print parse trace. It is left uninitialized so that multiple parsers can coexist. */ int yydebug; #else /* !YYDEBUG */ -# define YYDPRINTF(Args) ((void) 0) -# define YY_SYMBOL_PRINT(Title, Kind, Value, Location) -# define YY_STACK_PRINT(Bottom, Top) -# define YY_REDUCE_PRINT(Rule) +#define YYDPRINTF(Args) ((void)0) +#define YY_SYMBOL_PRINT(Title, Kind, Value, Location) +#define YY_STACK_PRINT(Bottom, Top) +#define YY_REDUCE_PRINT(Rule) #endif /* !YYDEBUG */ - /* YYINITDEPTH -- initial size of the parser's stacks. */ #ifndef YYINITDEPTH -# define YYINITDEPTH 200 +#define YYINITDEPTH 200 #endif /* YYMAXDEPTH -- maximum size the stacks can grow to (effective only @@ -1132,16 +3705,15 @@ int yydebug; evaluated with infinite-precision integer arithmetic. */ #ifndef YYMAXDEPTH -# define YYMAXDEPTH 10000 +#define YYMAXDEPTH 10000 #endif - /* Context of a parse error. */ typedef struct { - yy_state_t *yyssp; + yy_state_t *yyssp; yysymbol_kind_t yytoken; - YYLTYPE *yylloc; + YYLTYPE *yylloc; } yypcontext_t; /* Put in YYARG at most YYARGN of the expected tokens given the @@ -1150,69 +3722,59 @@ typedef struct be less than YYNTOKENS). Return YYENOMEM on memory exhaustion. Return 0 if there are more than YYARGN expected tokens, yet fill YYARG up to YYARGN. */ -static int -yypcontext_expected_tokens (const yypcontext_t *yyctx, - yysymbol_kind_t yyarg[], int yyargn) +static int yypcontext_expected_tokens(const yypcontext_t *yyctx, yysymbol_kind_t yyarg[], int yyargn) { /* Actual size of YYARG. */ int yycount = 0; - int yyn = yypact[+*yyctx->yyssp]; - if (!yypact_value_is_default (yyn)) - { - /* Start YYX at -YYN if negative to avoid negative indexes in - YYCHECK. In other words, skip the first -YYN actions for - this state because they are default actions. */ - int yyxbegin = yyn < 0 ? -yyn : 0; - /* Stay within bounds of both yycheck and yytname. */ - int yychecklim = YYLAST - yyn + 1; - int yyxend = yychecklim < YYNTOKENS ? yychecklim : YYNTOKENS; - int yyx; - for (yyx = yyxbegin; yyx < yyxend; ++yyx) - if (yycheck[yyx + yyn] == yyx && yyx != YYSYMBOL_YYerror - && !yytable_value_is_error (yytable[yyx + yyn])) - { - if (!yyarg) - ++yycount; - else if (yycount == yyargn) - return 0; - else - yyarg[yycount++] = YY_CAST (yysymbol_kind_t, yyx); - } - } + int yyn = yypact[+*yyctx->yyssp]; + if (!yypact_value_is_default(yyn)) { + /* Start YYX at -YYN if negative to avoid negative indexes in + YYCHECK. In other words, skip the first -YYN actions for + this state because they are default actions. */ + int yyxbegin = yyn < 0 ? -yyn : 0; + /* Stay within bounds of both yycheck and yytname. */ + int yychecklim = YYLAST - yyn + 1; + int yyxend = yychecklim < YYNTOKENS ? yychecklim : YYNTOKENS; + int yyx; + for (yyx = yyxbegin; yyx < yyxend; ++yyx) + if (yycheck[yyx + yyn] == yyx && yyx != YYSYMBOL_YYerror && !yytable_value_is_error(yytable[yyx + yyn])) { + if (!yyarg) + ++yycount; + else if (yycount == yyargn) + return 0; + else + yyarg[yycount++] = YY_CAST(yysymbol_kind_t, yyx); + } + } if (yyarg && yycount == 0 && 0 < yyargn) yyarg[0] = YYSYMBOL_YYEMPTY; return yycount; } - - - #ifndef yystrlen -# if defined __GLIBC__ && defined _STRING_H -# define yystrlen(S) (YY_CAST (YYPTRDIFF_T, strlen (S))) -# else +#if defined __GLIBC__ && defined _STRING_H +#define yystrlen(S) (YY_CAST(YYPTRDIFF_T, strlen(S))) +#else /* Return the length of YYSTR. */ -static YYPTRDIFF_T -yystrlen (const char *yystr) +static YYPTRDIFF_T yystrlen(const char *yystr) { YYPTRDIFF_T yylen; for (yylen = 0; yystr[yylen]; yylen++) continue; return yylen; } -# endif +#endif #endif #ifndef yystpcpy -# if defined __GLIBC__ && defined _STRING_H && defined _GNU_SOURCE -# define yystpcpy stpcpy -# else +#if defined __GLIBC__ && defined _STRING_H && defined _GNU_SOURCE +#define yystpcpy stpcpy +#else /* Copy YYSRC to YYDEST, returning the address of the terminating '\0' in YYDEST. */ -static char * -yystpcpy (char *yydest, const char *yysrc) +static char *yystpcpy(char *yydest, const char *yysrc) { - char *yyd = yydest; + char *yyd = yydest; const char *yys = yysrc; while ((*yyd++ = *yys++) != '\0') @@ -1220,7 +3782,7 @@ yystpcpy (char *yydest, const char *yysrc) return yyd - 1; } -# endif +#endif #endif #ifndef yytnamerr @@ -1231,52 +3793,45 @@ yystpcpy (char *yydest, const char *yysrc) backslash-backslash). YYSTR is taken from yytname. If YYRES is null, do not copy; instead, return the length of what the result would have been. */ -static YYPTRDIFF_T -yytnamerr (char *yyres, const char *yystr) +static YYPTRDIFF_T yytnamerr(char *yyres, const char *yystr) { - if (*yystr == '"') - { - YYPTRDIFF_T yyn = 0; - char const *yyp = yystr; - for (;;) - switch (*++yyp) - { - case '\'': - case ',': + if (*yystr == '"') { + YYPTRDIFF_T yyn = 0; + char const *yyp = yystr; + for (;;) + switch (*++yyp) { + case '\'': + case ',': goto do_not_strip_quotes; + + case '\\': + if (*++yyp != '\\') goto do_not_strip_quotes; - - case '\\': - if (*++yyp != '\\') - goto do_not_strip_quotes; - else - goto append; - - append: - default: - if (yyres) - yyres[yyn] = *yyp; - yyn++; - break; - - case '"': - if (yyres) - yyres[yyn] = '\0'; - return yyn; - } - do_not_strip_quotes: ; - } + else + goto append; + + append: + default: + if (yyres) + yyres[yyn] = *yyp; + yyn++; + break; + + case '"': + if (yyres) + yyres[yyn] = '\0'; + return yyn; + } + do_not_strip_quotes:; + } if (yyres) - return yystpcpy (yyres, yystr) - yyres; + return yystpcpy(yyres, yystr) - yyres; else - return yystrlen (yystr); + return yystrlen(yystr); } #endif - -static int -yy_syntax_error_arguments (const yypcontext_t *yyctx, - yysymbol_kind_t yyarg[], int yyargn) +static int yy_syntax_error_arguments(const yypcontext_t *yyctx, yysymbol_kind_t yyarg[], int yyargn) { /* Actual size of YYARG. */ int yycount = 0; @@ -1303,19 +3858,17 @@ yy_syntax_error_arguments (const yypcontext_t *yyctx, one exception: it will still contain any token that will not be accepted due to an error action in a later state. */ - if (yyctx->yytoken != YYSYMBOL_YYEMPTY) - { - int yyn; - if (yyarg) - yyarg[yycount] = yyctx->yytoken; - ++yycount; - yyn = yypcontext_expected_tokens (yyctx, - yyarg ? yyarg + 1 : yyarg, yyargn - 1); - if (yyn == YYENOMEM) - return YYENOMEM; - else - yycount += yyn; - } + if (yyctx->yytoken != YYSYMBOL_YYEMPTY) { + int yyn; + if (yyarg) + yyarg[yycount] = yyctx->yytoken; + ++yycount; + yyn = yypcontext_expected_tokens(yyctx, yyarg ? yyarg + 1 : yyarg, yyargn - 1); + if (yyn == YYENOMEM) + return YYENOMEM; + else + yycount += yyn; + } return yycount; } @@ -1327,11 +3880,12 @@ yy_syntax_error_arguments (const yypcontext_t *yyctx, not large enough to hold the message. In that case, also set *YYMSG_ALLOC to the required number of bytes. Return YYENOMEM if the required number of bytes is too large to store. */ -static int -yysyntax_error (YYPTRDIFF_T *yymsg_alloc, char **yymsg, - const yypcontext_t *yyctx) +static int yysyntax_error(YYPTRDIFF_T *yymsg_alloc, char **yymsg, const yypcontext_t *yyctx) { - enum { YYARGS_MAX = 5 }; + enum + { + YYARGS_MAX = 5 + }; /* Internationalized format string. */ const char *yyformat = YY_NULLPTR; /* Arguments of yyformat: reported tokens (one for the "unexpected", @@ -1341,16 +3895,13 @@ yysyntax_error (YYPTRDIFF_T *yymsg_alloc, char **yymsg, YYPTRDIFF_T yysize = 0; /* Actual size of YYARG. */ - int yycount = yy_syntax_error_arguments (yyctx, yyarg, YYARGS_MAX); + int yycount = yy_syntax_error_arguments(yyctx, yyarg, YYARGS_MAX); if (yycount == YYENOMEM) return YYENOMEM; - switch (yycount) - { -#define YYCASE_(N, S) \ - case N: \ - yyformat = S; \ - break + switch (yycount) { +#define YYCASE_(N, S) \ + case N: yyformat = S; break default: /* Avoid compiler warnings. */ YYCASE_(0, YY_("syntax error")); YYCASE_(1, YY_("syntax error, unexpected %s")); @@ -1359,134 +3910,118 @@ yysyntax_error (YYPTRDIFF_T *yymsg_alloc, char **yymsg, YYCASE_(4, YY_("syntax error, unexpected %s, expecting %s or %s or %s")); YYCASE_(5, YY_("syntax error, unexpected %s, expecting %s or %s or %s or %s")); #undef YYCASE_ - } + } /* Compute error message size. Don't count the "%s"s, but reserve room for the terminator. */ - yysize = yystrlen (yyformat) - 2 * yycount + 1; + yysize = yystrlen(yyformat) - 2 * yycount + 1; { int yyi; - for (yyi = 0; yyi < yycount; ++yyi) - { - YYPTRDIFF_T yysize1 - = yysize + yytnamerr (YY_NULLPTR, yytname[yyarg[yyi]]); - if (yysize <= yysize1 && yysize1 <= YYSTACK_ALLOC_MAXIMUM) - yysize = yysize1; - else - return YYENOMEM; - } + for (yyi = 0; yyi < yycount; ++yyi) { + YYPTRDIFF_T yysize1 = yysize + yytnamerr(YY_NULLPTR, yytname[yyarg[yyi]]); + if (yysize <= yysize1 && yysize1 <= YYSTACK_ALLOC_MAXIMUM) + yysize = yysize1; + else + return YYENOMEM; + } } - if (*yymsg_alloc < yysize) - { - *yymsg_alloc = 2 * yysize; - if (! (yysize <= *yymsg_alloc - && *yymsg_alloc <= YYSTACK_ALLOC_MAXIMUM)) - *yymsg_alloc = YYSTACK_ALLOC_MAXIMUM; - return -1; - } + if (*yymsg_alloc < yysize) { + *yymsg_alloc = 2 * yysize; + if (!(yysize <= *yymsg_alloc && *yymsg_alloc <= YYSTACK_ALLOC_MAXIMUM)) + *yymsg_alloc = YYSTACK_ALLOC_MAXIMUM; + return -1; + } /* Avoid sprintf, as that infringes on the user's name space. Don't have undefined behavior even if the translation produced a string with the wrong number of "%s"s. */ { char *yyp = *yymsg; - int yyi = 0; + int yyi = 0; while ((*yyp = *yyformat) != '\0') - if (*yyp == '%' && yyformat[1] == 's' && yyi < yycount) - { - yyp += yytnamerr (yyp, yytname[yyarg[yyi++]]); - yyformat += 2; - } - else - { - ++yyp; - ++yyformat; - } + if (*yyp == '%' && yyformat[1] == 's' && yyi < yycount) { + yyp += yytnamerr(yyp, yytname[yyarg[yyi++]]); + yyformat += 2; + } else { + ++yyp; + ++yyformat; + } } return 0; } - /*-----------------------------------------------. | Release the memory associated to this symbol. | `-----------------------------------------------*/ -static void -yydestruct (const char *yymsg, - yysymbol_kind_t yykind, YYSTYPE *yyvaluep, YYLTYPE *yylocationp, const char * sql_string, ParsedSqlResult * sql_result, void * scanner) +static void yydestruct(const char *yymsg, yysymbol_kind_t yykind, YYSTYPE *yyvaluep, YYLTYPE *yylocationp, + const char *sql_string, ParsedSqlResult *sql_result, void *scanner) { - YY_USE (yyvaluep); - YY_USE (yylocationp); - YY_USE (sql_string); - YY_USE (sql_result); - YY_USE (scanner); + YY_USE(yyvaluep); + YY_USE(yylocationp); + YY_USE(sql_string); + YY_USE(sql_result); + YY_USE(scanner); if (!yymsg) yymsg = "Deleting"; - YY_SYMBOL_PRINT (yymsg, yykind, yyvaluep, yylocationp); + YY_SYMBOL_PRINT(yymsg, yykind, yyvaluep, yylocationp); YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN - YY_USE (yykind); + YY_USE(yykind); YY_IGNORE_MAYBE_UNINITIALIZED_END } - - - - - /*----------. | yyparse. | `----------*/ -int -yyparse (const char * sql_string, ParsedSqlResult * sql_result, void * scanner) +int yyparse(const char *sql_string, ParsedSqlResult *sql_result, void *scanner) { -/* Lookahead token kind. */ -int yychar; - - -/* The semantic value of the lookahead symbol. */ -/* Default value used for initialization, for pacifying older GCCs - or non-GCC compilers. */ -YY_INITIAL_VALUE (static YYSTYPE yyval_default;) -YYSTYPE yylval YY_INITIAL_VALUE (= yyval_default); - -/* Location data for the lookahead symbol. */ -static YYLTYPE yyloc_default -# if defined YYLTYPE_IS_TRIVIAL && YYLTYPE_IS_TRIVIAL - = { 1, 1, 1, 1 } -# endif -; -YYLTYPE yylloc = yyloc_default; + /* Lookahead token kind. */ + int yychar; + + /* The semantic value of the lookahead symbol. */ + /* Default value used for initialization, for pacifying older GCCs + or non-GCC compilers. */ + YY_INITIAL_VALUE(static YYSTYPE yyval_default;) + YYSTYPE yylval YY_INITIAL_VALUE(= yyval_default); + + /* Location data for the lookahead symbol. */ + static YYLTYPE yyloc_default +#if defined YYLTYPE_IS_TRIVIAL && YYLTYPE_IS_TRIVIAL + = {1, 1, 1, 1} +#endif + ; + YYLTYPE yylloc = yyloc_default; - /* Number of syntax errors so far. */ - int yynerrs = 0; + /* Number of syntax errors so far. */ + int yynerrs = 0; - yy_state_fast_t yystate = 0; - /* Number of tokens to shift before error messages enabled. */ - int yyerrstatus = 0; + yy_state_fast_t yystate = 0; + /* Number of tokens to shift before error messages enabled. */ + int yyerrstatus = 0; - /* Refer to the stacks through separate pointers, to allow yyoverflow - to reallocate them elsewhere. */ + /* Refer to the stacks through separate pointers, to allow yyoverflow + to reallocate them elsewhere. */ - /* Their size. */ - YYPTRDIFF_T yystacksize = YYINITDEPTH; + /* Their size. */ + YYPTRDIFF_T yystacksize = YYINITDEPTH; - /* The state stack: array, bottom, top. */ - yy_state_t yyssa[YYINITDEPTH]; - yy_state_t *yyss = yyssa; - yy_state_t *yyssp = yyss; + /* The state stack: array, bottom, top. */ + yy_state_t yyssa[YYINITDEPTH]; + yy_state_t *yyss = yyssa; + yy_state_t *yyssp = yyss; - /* The semantic value stack: array, bottom, top. */ - YYSTYPE yyvsa[YYINITDEPTH]; - YYSTYPE *yyvs = yyvsa; - YYSTYPE *yyvsp = yyvs; + /* The semantic value stack: array, bottom, top. */ + YYSTYPE yyvsa[YYINITDEPTH]; + YYSTYPE *yyvs = yyvsa; + YYSTYPE *yyvsp = yyvs; - /* The location stack: array, bottom, top. */ - YYLTYPE yylsa[YYINITDEPTH]; - YYLTYPE *yyls = yylsa; - YYLTYPE *yylsp = yyls; + /* The location stack: array, bottom, top. */ + YYLTYPE yylsa[YYINITDEPTH]; + YYLTYPE *yyls = yylsa; + YYLTYPE *yylsp = yyls; int yyn; /* The return value of yyparse. */ @@ -1502,24 +4037,23 @@ YYLTYPE yylloc = yyloc_default; YYLTYPE yyerror_range[3]; /* Buffer for error messages, and its allocated size. */ - char yymsgbuf[128]; - char *yymsg = yymsgbuf; + char yymsgbuf[128]; + char *yymsg = yymsgbuf; YYPTRDIFF_T yymsg_alloc = sizeof yymsgbuf; -#define YYPOPSTACK(N) (yyvsp -= (N), yyssp -= (N), yylsp -= (N)) +#define YYPOPSTACK(N) (yyvsp -= (N), yyssp -= (N), yylsp -= (N)) /* The number of symbols on the RHS of the reduced rule. Keep to zero when no symbol should be popped. */ int yylen = 0; - YYDPRINTF ((stderr, "Starting parse\n")); + YYDPRINTF((stderr, "Starting parse\n")); yychar = YYEMPTY; /* Cause a token to be read. */ yylsp[0] = yylloc; goto yysetstate; - /*------------------------------------------------------------. | yynewstate -- push a new state, which is found in yystate. | `------------------------------------------------------------*/ @@ -1528,93 +4062,90 @@ YYLTYPE yylloc = yyloc_default; have just been pushed. So pushing a state here evens the stacks. */ yyssp++; - /*--------------------------------------------------------------------. | yysetstate -- set current state (the top of the stack) to yystate. | `--------------------------------------------------------------------*/ yysetstate: - YYDPRINTF ((stderr, "Entering state %d\n", yystate)); - YY_ASSERT (0 <= yystate && yystate < YYNSTATES); + YYDPRINTF((stderr, "Entering state %d\n", yystate)); + YY_ASSERT(0 <= yystate && yystate < YYNSTATES); YY_IGNORE_USELESS_CAST_BEGIN - *yyssp = YY_CAST (yy_state_t, yystate); + *yyssp = YY_CAST(yy_state_t, yystate); YY_IGNORE_USELESS_CAST_END - YY_STACK_PRINT (yyss, yyssp); + YY_STACK_PRINT(yyss, yyssp); if (yyss + yystacksize - 1 <= yyssp) #if !defined yyoverflow && !defined YYSTACK_RELOCATE YYNOMEM; #else + { + /* Get the current used size of the three stacks, in elements. */ + YYPTRDIFF_T yysize = yyssp - yyss + 1; + +#if defined yyoverflow { - /* Get the current used size of the three stacks, in elements. */ - YYPTRDIFF_T yysize = yyssp - yyss + 1; - -# if defined yyoverflow - { - /* Give user a chance to reallocate the stack. Use copies of - these so that the &'s don't force the real ones into - memory. */ - yy_state_t *yyss1 = yyss; - YYSTYPE *yyvs1 = yyvs; - YYLTYPE *yyls1 = yyls; - - /* Each stack pointer address is followed by the size of the - data in use in that stack, in bytes. This used to be a - conditional around just the two extra args, but that might - be undefined if yyoverflow is a macro. */ - yyoverflow (YY_("memory exhausted"), - &yyss1, yysize * YYSIZEOF (*yyssp), - &yyvs1, yysize * YYSIZEOF (*yyvsp), - &yyls1, yysize * YYSIZEOF (*yylsp), - &yystacksize); - yyss = yyss1; - yyvs = yyvs1; - yyls = yyls1; - } -# else /* defined YYSTACK_RELOCATE */ - /* Extend the stack our own way. */ - if (YYMAXDEPTH <= yystacksize) + /* Give user a chance to reallocate the stack. Use copies of + these so that the &'s don't force the real ones into + memory. */ + yy_state_t *yyss1 = yyss; + YYSTYPE *yyvs1 = yyvs; + YYLTYPE *yyls1 = yyls; + + /* Each stack pointer address is followed by the size of the + data in use in that stack, in bytes. This used to be a + conditional around just the two extra args, but that might + be undefined if yyoverflow is a macro. */ + yyoverflow(YY_("memory exhausted"), + &yyss1, + yysize * YYSIZEOF(*yyssp), + &yyvs1, + yysize * YYSIZEOF(*yyvsp), + &yyls1, + yysize * YYSIZEOF(*yylsp), + &yystacksize); + yyss = yyss1; + yyvs = yyvs1; + yyls = yyls1; + } +#else /* defined YYSTACK_RELOCATE */ + /* Extend the stack our own way. */ + if (YYMAXDEPTH <= yystacksize) + YYNOMEM; + yystacksize *= 2; + if (YYMAXDEPTH < yystacksize) + yystacksize = YYMAXDEPTH; + + { + yy_state_t *yyss1 = yyss; + union yyalloc *yyptr = YY_CAST(union yyalloc *, YYSTACK_ALLOC(YY_CAST(YYSIZE_T, YYSTACK_BYTES(yystacksize)))); + if (!yyptr) YYNOMEM; - yystacksize *= 2; - if (YYMAXDEPTH < yystacksize) - yystacksize = YYMAXDEPTH; - - { - yy_state_t *yyss1 = yyss; - union yyalloc *yyptr = - YY_CAST (union yyalloc *, - YYSTACK_ALLOC (YY_CAST (YYSIZE_T, YYSTACK_BYTES (yystacksize)))); - if (! yyptr) - YYNOMEM; - YYSTACK_RELOCATE (yyss_alloc, yyss); - YYSTACK_RELOCATE (yyvs_alloc, yyvs); - YYSTACK_RELOCATE (yyls_alloc, yyls); -# undef YYSTACK_RELOCATE - if (yyss1 != yyssa) - YYSTACK_FREE (yyss1); - } -# endif + YYSTACK_RELOCATE(yyss_alloc, yyss); + YYSTACK_RELOCATE(yyvs_alloc, yyvs); + YYSTACK_RELOCATE(yyls_alloc, yyls); +#undef YYSTACK_RELOCATE + if (yyss1 != yyssa) + YYSTACK_FREE(yyss1); + } +#endif - yyssp = yyss + yysize - 1; - yyvsp = yyvs + yysize - 1; - yylsp = yyls + yysize - 1; + yyssp = yyss + yysize - 1; + yyvsp = yyvs + yysize - 1; + yylsp = yyls + yysize - 1; - YY_IGNORE_USELESS_CAST_BEGIN - YYDPRINTF ((stderr, "Stack size increased to %ld\n", - YY_CAST (long, yystacksize))); - YY_IGNORE_USELESS_CAST_END + YY_IGNORE_USELESS_CAST_BEGIN + YYDPRINTF((stderr, "Stack size increased to %ld\n", YY_CAST(long, yystacksize))); + YY_IGNORE_USELESS_CAST_END - if (yyss + yystacksize - 1 <= yyssp) - YYABORT; - } + if (yyss + yystacksize - 1 <= yyssp) + YYABORT; + } #endif /* !defined yyoverflow && !defined YYSTACK_RELOCATE */ - if (yystate == YYFINAL) YYACCEPT; goto yybackup; - /*-----------. | yybackup. | `-----------*/ @@ -1624,40 +4155,34 @@ YYLTYPE yylloc = yyloc_default; /* First try to decide what to do without reference to lookahead token. */ yyn = yypact[yystate]; - if (yypact_value_is_default (yyn)) + if (yypact_value_is_default(yyn)) goto yydefault; /* Not known => get a lookahead token if don't already have one. */ /* YYCHAR is either empty, or end-of-input, or a valid lookahead. */ - if (yychar == YYEMPTY) - { - YYDPRINTF ((stderr, "Reading a token\n")); - yychar = yylex (&yylval, &yylloc, scanner); - } + if (yychar == YYEMPTY) { + YYDPRINTF((stderr, "Reading a token\n")); + yychar = yylex(&yylval, &yylloc, scanner); + } - if (yychar <= YYEOF) - { - yychar = YYEOF; - yytoken = YYSYMBOL_YYEOF; - YYDPRINTF ((stderr, "Now at end of input.\n")); - } - else if (yychar == YYerror) - { - /* The scanner already issued an error message, process directly - to error recovery. But do not keep the error token as - lookahead, it is too special and may lead us to an endless - loop in error recovery. */ - yychar = YYUNDEF; - yytoken = YYSYMBOL_YYerror; - yyerror_range[1] = yylloc; - goto yyerrlab1; - } - else - { - yytoken = YYTRANSLATE (yychar); - YY_SYMBOL_PRINT ("Next token is", yytoken, &yylval, &yylloc); - } + if (yychar <= YYEOF) { + yychar = YYEOF; + yytoken = YYSYMBOL_YYEOF; + YYDPRINTF((stderr, "Now at end of input.\n")); + } else if (yychar == YYerror) { + /* The scanner already issued an error message, process directly + to error recovery. But do not keep the error token as + lookahead, it is too special and may lead us to an endless + loop in error recovery. */ + yychar = YYUNDEF; + yytoken = YYSYMBOL_YYerror; + yyerror_range[1] = yylloc; + goto yyerrlab1; + } else { + yytoken = YYTRANSLATE(yychar); + YY_SYMBOL_PRINT("Next token is", yytoken, &yylval, &yylloc); + } /* If the proper action on seeing token YYTOKEN is to reduce or to detect an error, take that action. */ @@ -1665,13 +4190,12 @@ YYLTYPE yylloc = yyloc_default; if (yyn < 0 || YYLAST < yyn || yycheck[yyn] != yytoken) goto yydefault; yyn = yytable[yyn]; - if (yyn <= 0) - { - if (yytable_value_is_error (yyn)) - goto yyerrlab; - yyn = -yyn; - goto yyreduce; - } + if (yyn <= 0) { + if (yytable_value_is_error(yyn)) + goto yyerrlab; + yyn = -yyn; + goto yyreduce; + } /* Count tokens shifted since error; after three, turn off error status. */ @@ -1679,7 +4203,7 @@ YYLTYPE yylloc = yyloc_default; yyerrstatus--; /* Shift the lookahead token. */ - YY_SYMBOL_PRINT ("Shifting", yytoken, &yylval, &yylloc); + YY_SYMBOL_PRINT("Shifting", yytoken, &yylval, &yylloc); yystate = yyn; YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN *++yyvsp = yylval; @@ -1690,7 +4214,6 @@ YYLTYPE yylloc = yyloc_default; yychar = YYEMPTY; goto yynewstate; - /*-----------------------------------------------------------. | yydefault -- do the default action for the current state. | `-----------------------------------------------------------*/ @@ -1700,7 +4223,6 @@ YYLTYPE yylloc = yyloc_default; goto yyerrlab; goto yyreduce; - /*-----------------------------. | yyreduce -- do a reduction. | `-----------------------------*/ @@ -1716,360 +4238,832 @@ YYLTYPE yylloc = yyloc_default; users should not rely upon it. Assigning to YYVAL unconditionally makes the parser a bit smaller, and it avoids a GCC warning that YYVAL may be used uninitialized. */ - yyval = yyvsp[1-yylen]; + yyval = yyvsp[1 - yylen]; /* Default location. */ - YYLLOC_DEFAULT (yyloc, (yylsp - yylen), yylen); + YYLLOC_DEFAULT(yyloc, (yylsp - yylen), yylen); yyerror_range[1] = yyloc; - YY_REDUCE_PRINT (yyn); - switch (yyn) + YY_REDUCE_PRINT(yyn); + switch (yyn) { + case 2: /* commands: command_wrapper opt_semicolon */ +#line 283 "yacc_sql.y" { - case 2: /* commands: command_wrapper opt_semicolon */ -#line 190 "yacc_sql.y" - { - std::unique_ptr sql_node = std::unique_ptr((yyvsp[-1].sql_node)); - sql_result->add_sql_node(std::move(sql_node)); - } -#line 1734 "yacc_sql.cpp" + std::unique_ptr sql_node = std::unique_ptr((yyvsp[-1].sql_node)); + sql_result->add_sql_node(std::move(sql_node)); + } +#line 1936 "yacc_sql.cpp" break; - case 23: /* exit_stmt: EXIT */ -#line 220 "yacc_sql.y" - { + case 25: /* exit_stmt: EXIT */ +#line 315 "yacc_sql.y" + { (void)yynerrs; // 这么写为了消除yynerrs未使用的告警。如果你有更好的方法欢迎提PR (yyval.sql_node) = new ParsedSqlNode(SCF_EXIT); } -#line 1743 "yacc_sql.cpp" +#line 1945 "yacc_sql.cpp" break; - case 24: /* help_stmt: HELP */ -#line 226 "yacc_sql.y" - { + case 26: /* help_stmt: HELP */ +#line 321 "yacc_sql.y" + { (yyval.sql_node) = new ParsedSqlNode(SCF_HELP); } -#line 1751 "yacc_sql.cpp" +#line 1953 "yacc_sql.cpp" break; - case 25: /* sync_stmt: SYNC */ -#line 231 "yacc_sql.y" - { + case 27: /* sync_stmt: SYNC */ +#line 326 "yacc_sql.y" + { (yyval.sql_node) = new ParsedSqlNode(SCF_SYNC); } -#line 1759 "yacc_sql.cpp" +#line 1961 "yacc_sql.cpp" break; - case 26: /* begin_stmt: TRX_BEGIN */ -#line 237 "yacc_sql.y" - { + case 28: /* begin_stmt: TRX_BEGIN */ +#line 332 "yacc_sql.y" + { (yyval.sql_node) = new ParsedSqlNode(SCF_BEGIN); } -#line 1767 "yacc_sql.cpp" +#line 1969 "yacc_sql.cpp" break; - case 27: /* commit_stmt: TRX_COMMIT */ -#line 243 "yacc_sql.y" - { + case 29: /* commit_stmt: TRX_COMMIT */ +#line 338 "yacc_sql.y" + { (yyval.sql_node) = new ParsedSqlNode(SCF_COMMIT); } -#line 1775 "yacc_sql.cpp" +#line 1977 "yacc_sql.cpp" break; - case 28: /* rollback_stmt: TRX_ROLLBACK */ -#line 249 "yacc_sql.y" - { + case 30: /* rollback_stmt: TRX_ROLLBACK */ +#line 344 "yacc_sql.y" + { (yyval.sql_node) = new ParsedSqlNode(SCF_ROLLBACK); } -#line 1783 "yacc_sql.cpp" +#line 1985 "yacc_sql.cpp" break; - case 29: /* drop_table_stmt: DROP TABLE ID */ -#line 255 "yacc_sql.y" - { - (yyval.sql_node) = new ParsedSqlNode(SCF_DROP_TABLE); + case 31: /* drop_table_stmt: DROP TABLE ID */ +#line 350 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_DROP_TABLE); (yyval.sql_node)->drop_table.relation_name = (yyvsp[0].string); free((yyvsp[0].string)); } -#line 1793 "yacc_sql.cpp" +#line 1995 "yacc_sql.cpp" break; - case 30: /* show_tables_stmt: SHOW TABLES */ -#line 262 "yacc_sql.y" - { + case 32: /* show_tables_stmt: SHOW TABLES */ +#line 357 "yacc_sql.y" + { (yyval.sql_node) = new ParsedSqlNode(SCF_SHOW_TABLES); } -#line 1801 "yacc_sql.cpp" +#line 2003 "yacc_sql.cpp" break; - case 31: /* desc_table_stmt: DESC ID */ -#line 268 "yacc_sql.y" - { - (yyval.sql_node) = new ParsedSqlNode(SCF_DESC_TABLE); + case 33: /* desc_table_stmt: DESC ID */ +#line 363 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_DESC_TABLE); (yyval.sql_node)->desc_table.relation_name = (yyvsp[0].string); free((yyvsp[0].string)); } -#line 1811 "yacc_sql.cpp" +#line 2013 "yacc_sql.cpp" break; - case 32: /* create_index_stmt: CREATE INDEX ID ON ID LBRACE ID RBRACE */ -#line 277 "yacc_sql.y" + case 34: /* show_index_stmt: SHOW INDEX FROM relation */ +#line 372 "yacc_sql.y" { - (yyval.sql_node) = new ParsedSqlNode(SCF_CREATE_INDEX); - CreateIndexSqlNode &create_index = (yyval.sql_node)->create_index; - create_index.index_name = (yyvsp[-5].string); - create_index.relation_name = (yyvsp[-3].string); - create_index.attribute_name = (yyvsp[-1].string); - free((yyvsp[-5].string)); - free((yyvsp[-3].string)); - free((yyvsp[-1].string)); + (yyval.sql_node) = new ParsedSqlNode(SCF_SHOW_INDEX); + ShowIndexSqlNode &show_index = (yyval.sql_node)->show_index; + show_index.relation_name = (yyvsp[0].string); + free((yyvsp[0].string)); } -#line 1826 "yacc_sql.cpp" +#line 2024 "yacc_sql.cpp" break; - case 33: /* drop_index_stmt: DROP INDEX ID ON ID */ -#line 291 "yacc_sql.y" + case 35: /* create_index_stmt: CREATE opt_unique INDEX ID ON ID LBRACE attr_list RBRACE */ +#line 382 "yacc_sql.y" { - (yyval.sql_node) = new ParsedSqlNode(SCF_DROP_INDEX); - (yyval.sql_node)->drop_index.index_name = (yyvsp[-2].string); - (yyval.sql_node)->drop_index.relation_name = (yyvsp[0].string); - free((yyvsp[-2].string)); - free((yyvsp[0].string)); + (yyval.sql_node) = new ParsedSqlNode(SCF_CREATE_INDEX); + CreateIndexSqlNode &create_index = (yyval.sql_node)->create_index; + create_index.unique = (yyvsp[-7].unique); // 用 opt_unique 的返回值来确定是否 UNIQUE + create_index.index_name = (yyvsp[-5].string); + create_index.relation_name = (yyvsp[-3].string); + create_index.attribute_name.swap(*(yyvsp[-1].index_attr_list)); // $8 是 vector 类型 + delete (yyvsp[-1].index_attr_list); // 释放指针 + free((yyvsp[-5].string)); + free((yyvsp[-3].string)); } -#line 1838 "yacc_sql.cpp" +#line 2040 "yacc_sql.cpp" break; - case 34: /* create_table_stmt: CREATE TABLE ID LBRACE attr_def attr_def_list RBRACE storage_format */ -#line 301 "yacc_sql.y" + case 36: /* create_index_stmt: CREATE VECTOR_T INDEX ID ON ID LBRACE attr_list RBRACE WITH vector_index_config */ +#line 394 "yacc_sql.y" { - (yyval.sql_node) = new ParsedSqlNode(SCF_CREATE_TABLE); - CreateTableSqlNode &create_table = (yyval.sql_node)->create_table; - create_table.relation_name = (yyvsp[-5].string); + (yyval.sql_node) = new ParsedSqlNode(SCF_CREATE_INDEX); + CreateIndexSqlNode &create_index = (yyval.sql_node)->create_index; + create_index.unique = false; // 向量索引不支持 + create_index.index_name = (yyvsp[-7].string); + create_index.relation_name = (yyvsp[-5].string); + create_index.attribute_name.swap(*(yyvsp[-3].index_attr_list)); // $8 是 vector 类型 + create_index.vector_index_config = std::move(*(yyvsp[0].vector_index_config)); + delete (yyvsp[-3].index_attr_list); // 释放指针 + free((yyvsp[-7].string)); free((yyvsp[-5].string)); - - std::vector *src_attrs = (yyvsp[-2].attr_infos); - - if (src_attrs != nullptr) { - create_table.attr_infos.swap(*src_attrs); - delete src_attrs; - } - create_table.attr_infos.emplace_back(*(yyvsp[-3].attr_info)); - std::reverse(create_table.attr_infos.begin(), create_table.attr_infos.end()); - delete (yyvsp[-3].attr_info); - if ((yyvsp[0].string) != nullptr) { - create_table.storage_format = (yyvsp[0].string); - free((yyvsp[0].string)); - } } -#line 1863 "yacc_sql.cpp" +#line 2057 "yacc_sql.cpp" break; - case 35: /* attr_def_list: %empty */ -#line 324 "yacc_sql.y" + case 37: /* opt_unique: UNIQUE */ +#line 409 "yacc_sql.y" { - (yyval.attr_infos) = nullptr; + (yyval.unique) = true; } -#line 1871 "yacc_sql.cpp" +#line 2063 "yacc_sql.cpp" break; - case 36: /* attr_def_list: COMMA attr_def attr_def_list */ -#line 328 "yacc_sql.y" + case 38: /* opt_unique: %empty */ +#line 410 "yacc_sql.y" { - if ((yyvsp[0].attr_infos) != nullptr) { - (yyval.attr_infos) = (yyvsp[0].attr_infos); - } else { - (yyval.attr_infos) = new std::vector; - } - (yyval.attr_infos)->emplace_back(*(yyvsp[-1].attr_info)); - delete (yyvsp[-1].attr_info); + (yyval.unique) = false; } -#line 1885 "yacc_sql.cpp" +#line 2069 "yacc_sql.cpp" break; - case 37: /* attr_def: ID type LBRACE number RBRACE */ -#line 341 "yacc_sql.y" + case 39: /* index_type: IVFFLAT */ +#line 415 "yacc_sql.y" { - (yyval.attr_info) = new AttrInfoSqlNode; - (yyval.attr_info)->type = (AttrType)(yyvsp[-3].number); - (yyval.attr_info)->name = (yyvsp[-4].string); - (yyval.attr_info)->length = (yyvsp[-1].number); - free((yyvsp[-4].string)); + (yyval.index_type) = IndexType::VectorIVFFlatIndex; } -#line 1897 "yacc_sql.cpp" +#line 2077 "yacc_sql.cpp" break; - case 38: /* attr_def: ID type */ -#line 349 "yacc_sql.y" + case 40: /* vector_index_config: LBRACE DISTANCE EQ ID COMMA TYPE EQ index_type RBRACE */ +#line 422 "yacc_sql.y" { - (yyval.attr_info) = new AttrInfoSqlNode; - (yyval.attr_info)->type = (AttrType)(yyvsp[0].number); - (yyval.attr_info)->name = (yyvsp[-1].string); - (yyval.attr_info)->length = 4; - free((yyvsp[-1].string)); + (yyval.vector_index_config) = new VectorIndexConfig; + (yyval.vector_index_config)->distance_fn = (yyvsp[-5].string); + (yyval.vector_index_config)->index_type = (yyvsp[-1].index_type); + free((yyvsp[-5].string)); } -#line 1909 "yacc_sql.cpp" +#line 2088 "yacc_sql.cpp" break; - case 39: /* number: NUMBER */ -#line 358 "yacc_sql.y" - {(yyval.number) = (yyvsp[0].number);} -#line 1915 "yacc_sql.cpp" + case 41: /* vector_index_config: LBRACE DISTANCE EQ ID COMMA TYPE EQ index_type COMMA LISTS EQ value COMMA PROBES EQ + value RBRACE */ +#line 429 "yacc_sql.y" + { + (yyval.vector_index_config) = new VectorIndexConfig; + (yyval.vector_index_config)->distance_fn = (yyvsp[-13].string); + (yyval.vector_index_config)->index_type = (yyvsp[-9].index_type); + (yyval.vector_index_config)->lists = std::move(*(yyvsp[-5].value)); + (yyval.vector_index_config)->probes = std::move(*(yyvsp[-1].value)); + free((yyvsp[-13].string)); + } +#line 2101 "yacc_sql.cpp" break; - case 40: /* type: INT_T */ -#line 361 "yacc_sql.y" - { (yyval.number) = static_cast(AttrType::INTS); } -#line 1921 "yacc_sql.cpp" + case 42: /* vector_index_config: LBRACE TYPE EQ index_type COMMA DISTANCE EQ ID COMMA LISTS EQ value COMMA PROBES EQ + value RBRACE */ +#line 438 "yacc_sql.y" + { + (yyval.vector_index_config) = new VectorIndexConfig; + (yyval.vector_index_config)->distance_fn = (yyvsp[-9].string); + (yyval.vector_index_config)->index_type = (yyvsp[-13].index_type); + (yyval.vector_index_config)->lists = std::move(*(yyvsp[-5].value)); + (yyval.vector_index_config)->probes = std::move(*(yyvsp[-1].value)); + free((yyvsp[-9].string)); + } +#line 2114 "yacc_sql.cpp" break; - case 41: /* type: STRING_T */ -#line 362 "yacc_sql.y" - { (yyval.number) = static_cast(AttrType::CHARS); } -#line 1927 "yacc_sql.cpp" + case 43: /* attr_list: ID */ +#line 450 "yacc_sql.y" + { + (yyval.index_attr_list) = new std::vector; // 创建一个新的 vector + (yyval.index_attr_list)->emplace_back((yyvsp[0].string)); // 将列名加入 vector + free((yyvsp[0].string)); + } +#line 2124 "yacc_sql.cpp" break; - case 42: /* type: FLOAT_T */ -#line 363 "yacc_sql.y" - { (yyval.number) = static_cast(AttrType::FLOATS); } -#line 1933 "yacc_sql.cpp" + case 44: /* attr_list: ID COMMA attr_list */ +#line 456 "yacc_sql.y" + { + (yyval.index_attr_list) = (yyvsp[0].index_attr_list); // 使用现有的 vector + (yyval.index_attr_list) + ->emplace((yyval.index_attr_list)->begin(), (yyvsp[-2].string)); // 将新列名加入 vector 开头 + free((yyvsp[-2].string)); + } +#line 2134 "yacc_sql.cpp" break; - case 43: /* type: VECTOR_T */ -#line 364 "yacc_sql.y" - { (yyval.number) = static_cast(AttrType::VECTORS); } -#line 1939 "yacc_sql.cpp" + case 45: /* drop_index_stmt: DROP INDEX ID ON ID */ +#line 465 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_DROP_INDEX); + (yyval.sql_node)->drop_index.index_name = (yyvsp[-2].string); + (yyval.sql_node)->drop_index.relation_name = (yyvsp[0].string); + free((yyvsp[-2].string)); + free((yyvsp[0].string)); + } +#line 2146 "yacc_sql.cpp" break; - case 44: /* insert_stmt: INSERT INTO ID VALUES LBRACE value value_list RBRACE */ -#line 368 "yacc_sql.y" + case 46: /* create_table_stmt: CREATE TABLE ID LBRACE attr_def attr_def_list RBRACE storage_format AS select_stmt */ +#line 475 "yacc_sql.y" { - (yyval.sql_node) = new ParsedSqlNode(SCF_INSERT); - (yyval.sql_node)->insertion.relation_name = (yyvsp[-5].string); - if ((yyvsp[-1].value_list) != nullptr) { - (yyval.sql_node)->insertion.values.swap(*(yyvsp[-1].value_list)); - delete (yyvsp[-1].value_list); - } - (yyval.sql_node)->insertion.values.emplace_back(*(yyvsp[-2].value)); - std::reverse((yyval.sql_node)->insertion.values.begin(), (yyval.sql_node)->insertion.values.end()); - delete (yyvsp[-2].value); + (yyval.sql_node) = create_table_sql_node( + (yyvsp[-7].string), (yyvsp[-5].attr_info), (yyvsp[-4].attr_infos), (yyvsp[-2].string), (yyvsp[0].sql_node)); + } +#line 2154 "yacc_sql.cpp" + break; + + case 47: /* create_table_stmt: CREATE TABLE ID LBRACE attr_def attr_def_list RBRACE storage_format select_stmt */ +#line 479 "yacc_sql.y" + { + (yyval.sql_node) = create_table_sql_node( + (yyvsp[-6].string), (yyvsp[-4].attr_info), (yyvsp[-3].attr_infos), (yyvsp[-1].string), (yyvsp[0].sql_node)); + } +#line 2162 "yacc_sql.cpp" + break; + + case 48: /* create_table_stmt: CREATE TABLE ID LBRACE attr_def attr_def_list RBRACE storage_format */ +#line 483 "yacc_sql.y" + { + (yyval.sql_node) = create_table_sql_node( + (yyvsp[-5].string), (yyvsp[-3].attr_info), (yyvsp[-2].attr_infos), (yyvsp[0].string), nullptr); + } +#line 2170 "yacc_sql.cpp" + break; + + case 49: /* create_table_stmt: CREATE TABLE ID storage_format AS select_stmt */ +#line 487 "yacc_sql.y" + { + (yyval.sql_node) = + create_table_sql_node((yyvsp[-3].string), nullptr, nullptr, (yyvsp[-2].string), (yyvsp[0].sql_node)); + } +#line 2178 "yacc_sql.cpp" + break; + + case 50: /* create_table_stmt: CREATE TABLE ID storage_format select_stmt */ +#line 491 "yacc_sql.y" + { + (yyval.sql_node) = + create_table_sql_node((yyvsp[-2].string), nullptr, nullptr, (yyvsp[-1].string), (yyvsp[0].sql_node)); + } +#line 2186 "yacc_sql.cpp" + break; + + case 51: /* create_view_stmt: CREATE VIEW ID AS select_stmt */ +#line 498 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_CREATE_VIEW); + CreateViewSqlNode &create_view = (yyval.sql_node)->create_view; + create_view.relation_name = (yyvsp[-2].string); + create_view.create_view_select = std::make_unique(std::move((yyvsp[0].sql_node)->selection)); + free((yyvsp[-2].string)); + } +#line 2198 "yacc_sql.cpp" + break; + + case 52: /* create_view_stmt: CREATE VIEW ID LBRACE attr_list RBRACE AS select_stmt */ +#line 506 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_CREATE_VIEW); + CreateViewSqlNode &create_view = (yyval.sql_node)->create_view; + create_view.relation_name = (yyvsp[-5].string); + create_view.attribute_names = std::move(*(yyvsp[-3].index_attr_list)); + create_view.create_view_select = std::make_unique(std::move((yyvsp[0].sql_node)->selection)); free((yyvsp[-5].string)); } -#line 1956 "yacc_sql.cpp" +#line 2211 "yacc_sql.cpp" + break; + + case 53: /* drop_view_stmt: DROP VIEW ID */ +#line 518 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_DROP_VIEW); + (yyval.sql_node)->drop_view.relation_name = (yyvsp[0].string); + free((yyvsp[0].string)); + } +#line 2221 "yacc_sql.cpp" + break; + + case 54: /* attr_def_list: %empty */ +#line 527 "yacc_sql.y" + { + (yyval.attr_infos) = nullptr; + } +#line 2229 "yacc_sql.cpp" break; - case 45: /* value_list: %empty */ -#line 384 "yacc_sql.y" + case 55: /* attr_def_list: COMMA attr_def attr_def_list */ +#line 531 "yacc_sql.y" { - (yyval.value_list) = nullptr; + if ((yyvsp[0].attr_infos) != nullptr) { + (yyval.attr_infos) = (yyvsp[0].attr_infos); + } else { + (yyval.attr_infos) = new std::vector; + } + (yyval.attr_infos)->emplace_back(*(yyvsp[-1].attr_info)); + delete (yyvsp[-1].attr_info); } -#line 1964 "yacc_sql.cpp" +#line 2243 "yacc_sql.cpp" break; - case 46: /* value_list: COMMA value value_list */ -#line 387 "yacc_sql.y" - { - if ((yyvsp[0].value_list) != nullptr) { - (yyval.value_list) = (yyvsp[0].value_list); + case 56: /* attr_def: ID type LBRACE NUMBER RBRACE nullable_constraint */ +#line 544 "yacc_sql.y" + { + (yyval.attr_info) = new AttrInfoSqlNode; + (yyval.attr_info)->name = (yyvsp[-5].string); + (yyval.attr_info)->type = (AttrType)(yyvsp[-4].number); + if ((yyval.attr_info)->type == AttrType::CHARS) { + (yyval.attr_info)->length = (yyvsp[-2].number); + } else if ((yyval.attr_info)->type == AttrType::VECTORS) { + (yyval.attr_info)->length = sizeof(float) * (yyvsp[-2].number); } else { - (yyval.value_list) = new std::vector; + ASSERT(false, "$$->type is invalid."); + } + (yyval.attr_info)->nullable = (yyvsp[0].nullable_info); + if ((yyval.attr_info)->nullable) { + (yyval.attr_info)->length++; + } + free((yyvsp[-5].string)); + } +#line 2265 "yacc_sql.cpp" + break; + + case 57: /* attr_def: ID type nullable_constraint */ +#line 562 "yacc_sql.y" + { + (yyval.attr_info) = new AttrInfoSqlNode; + (yyval.attr_info)->type = (AttrType)(yyvsp[-1].number); + (yyval.attr_info)->name = (yyvsp[-2].string); + if ((yyval.attr_info)->type == AttrType::INTS) { + (yyval.attr_info)->length = sizeof(int); + } else if ((yyval.attr_info)->type == AttrType::FLOATS) { + (yyval.attr_info)->length = sizeof(float); + } else if ((yyval.attr_info)->type == AttrType::DATES) { + (yyval.attr_info)->length = sizeof(int); + } else if ((yyval.attr_info)->type == AttrType::CHARS) { + (yyval.attr_info)->length = sizeof(char); + } else if ((yyval.attr_info)->type == AttrType::VECTORS) { + (yyval.attr_info)->length = sizeof(float) * 1; + } else if ((yyval.attr_info)->type == AttrType::TEXTS) { + (yyval.attr_info)->length = 65535; + } else { + ASSERT(false, "$$->type is invalid."); + } + (yyval.attr_info)->nullable = (yyvsp[0].nullable_info); // 处理NULL/NOT NULL标记 + if ((yyval.attr_info)->nullable) { + (yyval.attr_info)->length++; + } + free((yyvsp[-2].string)); + } +#line 2295 "yacc_sql.cpp" + break; + + case 58: /* nullable_constraint: NOT NULL_T */ +#line 591 "yacc_sql.y" + { + (yyval.nullable_info) = false; // NOT NULL 对应的可空性为 false + } +#line 2303 "yacc_sql.cpp" + break; + + case 59: /* nullable_constraint: NULLABLE */ +#line 595 "yacc_sql.y" + { + (yyval.nullable_info) = true; // NULLABLE 对应的可空性为 true 2022 + } +#line 2311 "yacc_sql.cpp" + break; + + case 60: /* nullable_constraint: NULL_T */ +#line 599 "yacc_sql.y" + { + (yyval.nullable_info) = true; // NULL 对应的可空性也为 true 2023 + } +#line 2319 "yacc_sql.cpp" + break; + + case 61: /* nullable_constraint: %empty */ +#line 603 "yacc_sql.y" + { + (yyval.nullable_info) = true; // 默认情况为 NULL + } +#line 2327 "yacc_sql.cpp" + break; + + case 62: /* type: INT_T */ +#line 609 "yacc_sql.y" + { + (yyval.number) = static_cast(AttrType::INTS); + } +#line 2333 "yacc_sql.cpp" + break; + + case 63: /* type: STRING_T */ +#line 610 "yacc_sql.y" + { + (yyval.number) = static_cast(AttrType::CHARS); + } +#line 2339 "yacc_sql.cpp" + break; + + case 64: /* type: FLOAT_T */ +#line 611 "yacc_sql.y" + { + (yyval.number) = static_cast(AttrType::FLOATS); + } +#line 2345 "yacc_sql.cpp" + break; + + case 65: /* type: DATE_T */ +#line 612 "yacc_sql.y" + { + (yyval.number) = static_cast(AttrType::DATES); + } +#line 2351 "yacc_sql.cpp" + break; + + case 66: /* type: TEXT_T */ +#line 613 "yacc_sql.y" + { + (yyval.number) = static_cast(AttrType::TEXTS); + } +#line 2357 "yacc_sql.cpp" + break; + + case 67: /* type: VECTOR_T */ +#line 614 "yacc_sql.y" + { + (yyval.number) = static_cast(AttrType::VECTORS); + } +#line 2363 "yacc_sql.cpp" + break; + + case 68: /* insert_stmt: INSERT INTO ID VALUES values_list */ +#line 619 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_INSERT); + (yyval.sql_node)->insertion.relation_name = (yyvsp[-2].string); + if ((yyvsp[0].values_list) != nullptr) { + (yyval.sql_node)->insertion.values_list.swap(*(yyvsp[0].values_list)); + delete (yyvsp[0].values_list); } - (yyval.value_list)->emplace_back(*(yyvsp[-1].value)); - delete (yyvsp[-1].value); + free((yyvsp[-2].string)); } -#line 1978 "yacc_sql.cpp" +#line 2377 "yacc_sql.cpp" break; - case 47: /* value: NUMBER */ -#line 398 "yacc_sql.y" - { - (yyval.value) = new Value((int)(yyvsp[0].number)); - (yyloc) = (yylsp[0]); + case 69: /* insert_stmt: INSERT INTO ID LBRACE attr_list RBRACE VALUES values_list */ +#line 629 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_INSERT); + (yyval.sql_node)->insertion.relation_name = (yyvsp[-5].string); + (yyval.sql_node)->insertion.attr_names = std::move(*(yyvsp[-3].index_attr_list)); + if ((yyvsp[0].values_list) != nullptr) { + (yyval.sql_node)->insertion.values_list.swap(*(yyvsp[0].values_list)); + delete (yyvsp[0].values_list); + } + free((yyvsp[-5].string)); } -#line 1987 "yacc_sql.cpp" +#line 2392 "yacc_sql.cpp" break; - case 48: /* value: FLOAT */ -#line 402 "yacc_sql.y" - { - (yyval.value) = new Value((float)(yyvsp[0].floats)); - (yyloc) = (yylsp[0]); + case 70: /* values_list: LBRACE value_list RBRACE */ +#line 643 "yacc_sql.y" + { + (yyval.values_list) = new std::vector>; + (yyval.values_list)->emplace_back(*(yyvsp[-1].value_list)); + delete (yyvsp[-1].value_list); + } +#line 2402 "yacc_sql.cpp" + break; + + case 71: /* values_list: values_list COMMA LBRACE value_list RBRACE */ +#line 649 "yacc_sql.y" + { + (yyval.values_list)->emplace_back(*(yyvsp[-1].value_list)); + delete (yyvsp[-1].value_list); + } +#line 2411 "yacc_sql.cpp" + break; + + case 72: /* digits: NUMBER */ +#line 656 "yacc_sql.y" + { + (yyval.digits) = float((yyvsp[0].number)); } -#line 1996 "yacc_sql.cpp" +#line 2419 "yacc_sql.cpp" break; - case 49: /* value: SSS */ -#line 406 "yacc_sql.y" - { - char *tmp = common::substr((yyvsp[0].string),1,strlen((yyvsp[0].string))-2); + case 73: /* digits: '-' NUMBER */ +#line 660 "yacc_sql.y" + { + (yyval.digits) = float(-(yyvsp[0].number)); + } +#line 2427 "yacc_sql.cpp" + break; + + case 74: /* digits: FLOAT */ +#line 664 "yacc_sql.y" + { + (yyval.digits) = (yyvsp[0].floats); + } +#line 2435 "yacc_sql.cpp" + break; + + case 75: /* digits: '-' FLOAT */ +#line 668 "yacc_sql.y" + { + (yyval.digits) = (yyvsp[0].floats); + } +#line 2443 "yacc_sql.cpp" + break; + + case 76: /* digits_list: %empty */ +#line 675 "yacc_sql.y" + { + (yyval.digits_list) = new std::vector(); + } +#line 2451 "yacc_sql.cpp" + break; + + case 77: /* digits_list: digits */ +#line 679 "yacc_sql.y" + { + (yyval.digits_list) = new std::vector(); + (yyval.digits_list)->push_back((yyvsp[0].digits)); + } +#line 2460 "yacc_sql.cpp" + break; + + case 78: /* digits_list: digits_list COMMA digits */ +#line 684 "yacc_sql.y" + { + (yyval.digits_list)->push_back((yyvsp[0].digits)); + } +#line 2468 "yacc_sql.cpp" + break; + + case 79: /* value_list: %empty */ +#line 691 "yacc_sql.y" + { + (yyval.value_list) = new std::vector; + } +#line 2476 "yacc_sql.cpp" + break; + + case 80: /* value_list: value */ +#line 695 "yacc_sql.y" + { + (yyval.value_list) = new std::vector; + (yyval.value_list)->emplace_back(*(yyvsp[0].value)); + delete (yyvsp[0].value); + } +#line 2486 "yacc_sql.cpp" + break; + + case 81: /* value_list: value_list COMMA value */ +#line 701 "yacc_sql.y" + { + (yyval.value_list)->emplace_back(*(yyvsp[0].value)); + delete (yyvsp[0].value); + } +#line 2495 "yacc_sql.cpp" + break; + + case 82: /* value: nonnegative_value */ +#line 708 "yacc_sql.y" + { + (yyval.value) = (yyvsp[0].value); + } +#line 2503 "yacc_sql.cpp" + break; + + case 83: /* value: '-' NUMBER */ +#line 711 "yacc_sql.y" + { + (yyval.value) = new Value(-(yyvsp[0].number)); + (yyloc) = (yylsp[-1]); + } +#line 2512 "yacc_sql.cpp" + break; + + case 84: /* value: '-' FLOAT */ +#line 715 "yacc_sql.y" + { + (yyval.value) = new Value(-(yyvsp[0].floats)); + (yyloc) = (yylsp[-1]); + } +#line 2521 "yacc_sql.cpp" + break; + + case 85: /* nonnegative_value: NUMBER */ +#line 722 "yacc_sql.y" + { + (yyval.value) = new Value((yyvsp[0].number)); + (yyloc) = (yylsp[0]); + } +#line 2530 "yacc_sql.cpp" + break; + + case 86: /* nonnegative_value: FLOAT */ +#line 726 "yacc_sql.y" + { + (yyval.value) = new Value((yyvsp[0].floats)); + (yyloc) = (yylsp[0]); + } +#line 2539 "yacc_sql.cpp" + break; + + case 87: /* nonnegative_value: SSS */ +#line 730 "yacc_sql.y" + { + char *tmp = common::substr((yyvsp[0].string), 1, strlen((yyvsp[0].string)) - 2); (yyval.value) = new Value(tmp); free(tmp); free((yyvsp[0].string)); } -#line 2007 "yacc_sql.cpp" +#line 2550 "yacc_sql.cpp" break; - case 50: /* storage_format: %empty */ -#line 415 "yacc_sql.y" + case 88: /* nonnegative_value: TRUE */ +#line 736 "yacc_sql.y" + { + (yyval.value) = new Value(true); + } +#line 2558 "yacc_sql.cpp" + break; + + case 89: /* nonnegative_value: FALSE */ +#line 739 "yacc_sql.y" + { + (yyval.value) = new Value(false); + } +#line 2566 "yacc_sql.cpp" + break; + + case 90: /* nonnegative_value: NULL_T */ +#line 742 "yacc_sql.y" + { + (yyval.value) = new Value(NullValue()); + } +#line 2574 "yacc_sql.cpp" + break; + + case 91: /* nonnegative_value: LSBRACE digits_list RSBRACE */ +#line 745 "yacc_sql.y" + { + (yyval.value) = new Value(*(yyvsp[-1].digits_list)); + } +#line 2582 "yacc_sql.cpp" + break; + + case 92: /* storage_format: %empty */ +#line 752 "yacc_sql.y" { (yyval.string) = nullptr; } -#line 2015 "yacc_sql.cpp" +#line 2590 "yacc_sql.cpp" break; - case 51: /* storage_format: STORAGE FORMAT EQ ID */ -#line 419 "yacc_sql.y" + case 93: /* storage_format: STORAGE FORMAT EQ ID */ +#line 756 "yacc_sql.y" { (yyval.string) = (yyvsp[0].string); } -#line 2023 "yacc_sql.cpp" +#line 2598 "yacc_sql.cpp" break; - case 52: /* delete_stmt: DELETE FROM ID where */ -#line 426 "yacc_sql.y" + case 94: /* delete_stmt: DELETE FROM ID where */ +#line 763 "yacc_sql.y" { - (yyval.sql_node) = new ParsedSqlNode(SCF_DELETE); + (yyval.sql_node) = new ParsedSqlNode(SCF_DELETE); (yyval.sql_node)->deletion.relation_name = (yyvsp[-1].string); - if ((yyvsp[0].condition_list) != nullptr) { - (yyval.sql_node)->deletion.conditions.swap(*(yyvsp[0].condition_list)); - delete (yyvsp[0].condition_list); + if ((yyvsp[0].expression) != nullptr) { + (yyval.sql_node)->deletion.condition = std::unique_ptr((yyvsp[0].expression)); } free((yyvsp[-1].string)); } -#line 2037 "yacc_sql.cpp" +#line 2611 "yacc_sql.cpp" break; - case 53: /* update_stmt: UPDATE ID SET ID EQ value where */ -#line 438 "yacc_sql.y" + case 95: /* update_stmt: UPDATE ID SET set_clauses where */ +#line 775 "yacc_sql.y" { - (yyval.sql_node) = new ParsedSqlNode(SCF_UPDATE); - (yyval.sql_node)->update.relation_name = (yyvsp[-5].string); - (yyval.sql_node)->update.attribute_name = (yyvsp[-3].string); - (yyval.sql_node)->update.value = *(yyvsp[-1].value); - if ((yyvsp[0].condition_list) != nullptr) { - (yyval.sql_node)->update.conditions.swap(*(yyvsp[0].condition_list)); - delete (yyvsp[0].condition_list); + (yyval.sql_node) = new ParsedSqlNode(SCF_UPDATE); + (yyval.sql_node)->update.relation_name = (yyvsp[-3].string); + (yyval.sql_node)->update.set_clauses.swap(*(yyvsp[-1].set_clauses)); + if ((yyvsp[0].expression) != nullptr) { + (yyval.sql_node)->update.conditions = std::unique_ptr((yyvsp[0].expression)); } - free((yyvsp[-5].string)); free((yyvsp[-3].string)); + delete (yyvsp[-1].set_clauses); + } +#line 2626 "yacc_sql.cpp" + break; + + case 96: /* set_clauses: set_clause */ +#line 789 "yacc_sql.y" + { + (yyval.set_clauses) = new std::vector; + (yyval.set_clauses)->emplace_back(std::move(*(yyvsp[0].set_clause))); } -#line 2054 "yacc_sql.cpp" +#line 2635 "yacc_sql.cpp" break; - case 54: /* select_stmt: SELECT expression_list FROM rel_list where group_by */ -#line 453 "yacc_sql.y" + case 97: /* set_clauses: set_clauses COMMA set_clause */ +#line 794 "yacc_sql.y" + { + (yyval.set_clauses)->emplace_back(std::move(*(yyvsp[0].set_clause))); + } +#line 2643 "yacc_sql.cpp" + break; + + case 98: /* set_clause: ID EQ expression */ +#line 801 "yacc_sql.y" + { + (yyval.set_clause) = new SetClauseSqlNode; + (yyval.set_clause)->field_name = (yyvsp[-2].string); + (yyval.set_clause)->value = std::unique_ptr((yyvsp[0].expression)); + free((yyvsp[-2].string)); + } +#line 2654 "yacc_sql.cpp" + break; + + case 99: /* select_stmt: SELECT expression_list FROM rel_list where group_by opt_having opt_order_by opt_limit */ +#line 811 "yacc_sql.y" { (yyval.sql_node) = new ParsedSqlNode(SCF_SELECT); - if ((yyvsp[-4].expression_list) != nullptr) { - (yyval.sql_node)->selection.expressions.swap(*(yyvsp[-4].expression_list)); - delete (yyvsp[-4].expression_list); + if ((yyvsp[-7].expression_list) != nullptr) { + (yyval.sql_node)->selection.expressions.swap(*(yyvsp[-7].expression_list)); + delete (yyvsp[-7].expression_list); + } + + if ((yyvsp[-5].relation_list) != nullptr) { + (yyval.sql_node)->selection.relations.swap(*(yyvsp[-5].relation_list)); + delete (yyvsp[-5].relation_list); + } + + (yyval.sql_node)->selection.conditions = nullptr; + + if ((yyvsp[-4].expression) != nullptr) { + (yyval.sql_node)->selection.conditions = std::unique_ptr((yyvsp[-4].expression)); } - if ((yyvsp[-2].relation_list) != nullptr) { - (yyval.sql_node)->selection.relations.swap(*(yyvsp[-2].relation_list)); - delete (yyvsp[-2].relation_list); + if ((yyvsp[-3].expression_list) != nullptr) { + (yyval.sql_node)->selection.group_by.swap(*(yyvsp[-3].expression_list)); + delete (yyvsp[-3].expression_list); } - if ((yyvsp[-1].condition_list) != nullptr) { - (yyval.sql_node)->selection.conditions.swap(*(yyvsp[-1].condition_list)); - delete (yyvsp[-1].condition_list); + if ((yyvsp[-2].expression) != nullptr) { + (yyval.sql_node)->selection.having_conditions = std::unique_ptr((yyvsp[-2].expression)); + } + + if ((yyvsp[-1].orderby_list) != nullptr) { + (yyval.sql_node)->selection.order_by.swap(*(yyvsp[-1].orderby_list)); + delete (yyvsp[-1].orderby_list); + } + + if ((yyvsp[0].limited_info) != nullptr) { + (yyval.sql_node)->selection.limit = std::make_unique(*(yyvsp[0].limited_info)); + delete (yyvsp[0].limited_info); + } + } +#line 2696 "yacc_sql.cpp" + break; + + case 100: /* select_stmt: SELECT expression_list FROM relation INNER JOIN join_clauses where group_by */ +#line 849 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_SELECT); + if ((yyvsp[-7].expression_list) != nullptr) { + (yyval.sql_node)->selection.expressions.swap(*(yyvsp[-7].expression_list)); + delete (yyvsp[-7].expression_list); + } + + if ((yyvsp[-5].string) != nullptr) { + (yyval.sql_node)->selection.relations.emplace_back((yyvsp[-5].string)); + free((yyvsp[-5].string)); + } + + if ((yyvsp[-2].join_clauses) != nullptr) { + for (auto it = (yyvsp[-2].join_clauses)->relations.rbegin(); it != (yyvsp[-2].join_clauses)->relations.rend(); + ++it) { + (yyval.sql_node)->selection.relations.emplace_back(std::move(*it)); + } + (yyval.sql_node)->selection.conditions = std::move((yyvsp[-2].join_clauses)->conditions); + } + + if ((yyvsp[-1].expression) != nullptr) { + auto ptr = (yyval.sql_node)->selection.conditions.release(); + (yyval.sql_node)->selection.conditions = + std::make_unique(ConjunctionExpr::Type::AND, ptr, (yyvsp[-1].expression)); } if ((yyvsp[0].expression_list) != nullptr) { @@ -2077,366 +5071,612 @@ YYLTYPE yylloc = yyloc_default; delete (yyvsp[0].expression_list); } } -#line 2081 "yacc_sql.cpp" +#line 2730 "yacc_sql.cpp" break; - case 55: /* calc_stmt: CALC expression_list */ -#line 478 "yacc_sql.y" + case 101: /* calc_stmt: CALC expression_list */ +#line 882 "yacc_sql.y" { (yyval.sql_node) = new ParsedSqlNode(SCF_CALC); (yyval.sql_node)->calc.expressions.swap(*(yyvsp[0].expression_list)); delete (yyvsp[0].expression_list); } -#line 2091 "yacc_sql.cpp" +#line 2740 "yacc_sql.cpp" break; - case 56: /* expression_list: expression */ -#line 487 "yacc_sql.y" + case 102: /* calc_stmt: SELECT expression_list */ +#line 888 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_CALC); + (yyval.sql_node)->calc.expressions.swap(*(yyvsp[0].expression_list)); + delete (yyvsp[0].expression_list); + } +#line 2750 "yacc_sql.cpp" + break; + + case 103: /* expression_list: %empty */ +#line 896 "yacc_sql.y" { (yyval.expression_list) = new std::vector>; - (yyval.expression_list)->emplace_back((yyvsp[0].expression)); } -#line 2100 "yacc_sql.cpp" +#line 2758 "yacc_sql.cpp" break; - case 57: /* expression_list: expression COMMA expression_list */ -#line 492 "yacc_sql.y" + case 104: /* expression_list: expression alias */ +#line 900 "yacc_sql.y" + { + (yyval.expression_list) = new std::vector>; + if (nullptr != (yyvsp[0].string)) { + (yyvsp[-1].expression)->set_alias((yyvsp[0].string)); + } + (yyval.expression_list)->emplace_back((yyvsp[-1].expression)); + free((yyvsp[0].string)); + } +#line 2771 "yacc_sql.cpp" + break; + + case 105: /* expression_list: expression alias COMMA expression_list */ +#line 909 "yacc_sql.y" { if ((yyvsp[0].expression_list) != nullptr) { (yyval.expression_list) = (yyvsp[0].expression_list); } else { (yyval.expression_list) = new std::vector>; } - (yyval.expression_list)->emplace((yyval.expression_list)->begin(), (yyvsp[-2].expression)); + if (nullptr != (yyvsp[-2].string)) { + (yyvsp[-3].expression)->set_alias((yyvsp[-2].string)); + } + (yyval.expression_list)->emplace((yyval.expression_list)->begin(), std::move((yyvsp[-3].expression))); + free((yyvsp[-2].string)); } -#line 2113 "yacc_sql.cpp" +#line 2788 "yacc_sql.cpp" break; - case 58: /* expression: expression '+' expression */ -#line 502 "yacc_sql.y" - { - (yyval.expression) = create_arithmetic_expression(ArithmeticExpr::Type::ADD, (yyvsp[-2].expression), (yyvsp[0].expression), sql_string, &(yyloc)); + case 106: /* expression: expression '+' expression */ +#line 924 "yacc_sql.y" + { + (yyval.expression) = create_arithmetic_expression( + ArithmeticExpr::Type::ADD, (yyvsp[-2].expression), (yyvsp[0].expression), sql_string, &(yyloc)); } -#line 2121 "yacc_sql.cpp" +#line 2796 "yacc_sql.cpp" break; - case 59: /* expression: expression '-' expression */ -#line 505 "yacc_sql.y" - { - (yyval.expression) = create_arithmetic_expression(ArithmeticExpr::Type::SUB, (yyvsp[-2].expression), (yyvsp[0].expression), sql_string, &(yyloc)); + case 107: /* expression: expression '-' expression */ +#line 927 "yacc_sql.y" + { + (yyval.expression) = create_arithmetic_expression( + ArithmeticExpr::Type::SUB, (yyvsp[-2].expression), (yyvsp[0].expression), sql_string, &(yyloc)); } -#line 2129 "yacc_sql.cpp" +#line 2804 "yacc_sql.cpp" break; - case 60: /* expression: expression '*' expression */ -#line 508 "yacc_sql.y" - { - (yyval.expression) = create_arithmetic_expression(ArithmeticExpr::Type::MUL, (yyvsp[-2].expression), (yyvsp[0].expression), sql_string, &(yyloc)); + case 108: /* expression: expression '*' expression */ +#line 930 "yacc_sql.y" + { + (yyval.expression) = create_arithmetic_expression( + ArithmeticExpr::Type::MUL, (yyvsp[-2].expression), (yyvsp[0].expression), sql_string, &(yyloc)); } -#line 2137 "yacc_sql.cpp" +#line 2812 "yacc_sql.cpp" break; - case 61: /* expression: expression '/' expression */ -#line 511 "yacc_sql.y" - { - (yyval.expression) = create_arithmetic_expression(ArithmeticExpr::Type::DIV, (yyvsp[-2].expression), (yyvsp[0].expression), sql_string, &(yyloc)); + case 109: /* expression: expression '/' expression */ +#line 933 "yacc_sql.y" + { + (yyval.expression) = create_arithmetic_expression( + ArithmeticExpr::Type::DIV, (yyvsp[-2].expression), (yyvsp[0].expression), sql_string, &(yyloc)); } -#line 2145 "yacc_sql.cpp" +#line 2820 "yacc_sql.cpp" break; - case 62: /* expression: LBRACE expression RBRACE */ -#line 514 "yacc_sql.y" - { - (yyval.expression) = (yyvsp[-1].expression); + case 110: /* expression: LBRACE expression_list RBRACE */ +#line 936 "yacc_sql.y" + { + if ((yyvsp[-1].expression_list)->size() == 1) { + (yyval.expression) = (yyvsp[-1].expression_list)->front().get(); + } else { + (yyval.expression) = new ListExpr(std::move(*(yyvsp[-1].expression_list))); + } (yyval.expression)->set_name(token_name(sql_string, &(yyloc))); } -#line 2154 "yacc_sql.cpp" +#line 2833 "yacc_sql.cpp" break; - case 63: /* expression: '-' expression */ -#line 518 "yacc_sql.y" - { - (yyval.expression) = create_arithmetic_expression(ArithmeticExpr::Type::NEGATIVE, (yyvsp[0].expression), nullptr, sql_string, &(yyloc)); + case 111: /* expression: '-' expression */ +#line 944 "yacc_sql.y" + { + (yyval.expression) = create_arithmetic_expression( + ArithmeticExpr::Type::NEGATIVE, (yyvsp[0].expression), nullptr, sql_string, &(yyloc)); } -#line 2162 "yacc_sql.cpp" +#line 2841 "yacc_sql.cpp" break; - case 64: /* expression: value */ -#line 521 "yacc_sql.y" - { + case 112: /* expression: nonnegative_value */ +#line 947 "yacc_sql.y" + { (yyval.expression) = new ValueExpr(*(yyvsp[0].value)); (yyval.expression)->set_name(token_name(sql_string, &(yyloc))); delete (yyvsp[0].value); } -#line 2172 "yacc_sql.cpp" +#line 2851 "yacc_sql.cpp" break; - case 65: /* expression: rel_attr */ -#line 526 "yacc_sql.y" - { + case 113: /* expression: rel_attr */ +#line 952 "yacc_sql.y" + { RelAttrSqlNode *node = (yyvsp[0].rel_attr); - (yyval.expression) = new UnboundFieldExpr(node->relation_name, node->attribute_name); + (yyval.expression) = new UnboundFieldExpr(node->relation_name, node->attribute_name); (yyval.expression)->set_name(token_name(sql_string, &(yyloc))); delete (yyvsp[0].rel_attr); } -#line 2183 "yacc_sql.cpp" +#line 2862 "yacc_sql.cpp" break; - case 66: /* expression: '*' */ -#line 532 "yacc_sql.y" - { + case 114: /* expression: '*' */ +#line 958 "yacc_sql.y" + { (yyval.expression) = new StarExpr(); } -#line 2191 "yacc_sql.cpp" +#line 2870 "yacc_sql.cpp" + break; + + case 115: /* expression: ID DOT '*' */ +#line 961 "yacc_sql.y" + { + (yyval.expression) = new StarExpr((yyvsp[-2].string)); + } +#line 2878 "yacc_sql.cpp" break; - case 67: /* rel_attr: ID */ -#line 539 "yacc_sql.y" - { - (yyval.rel_attr) = new RelAttrSqlNode; + case 116: /* expression: func_expr */ +#line 964 "yacc_sql.y" + { + (yyval.expression) = (yyvsp[0].expression); // AggrFuncExpr + } +#line 2886 "yacc_sql.cpp" + break; + + case 117: /* expression: sub_query_expr */ +#line 967 "yacc_sql.y" + { + (yyval.expression) = (yyvsp[0].expression); // SubQueryExpr + } +#line 2894 "yacc_sql.cpp" + break; + + case 118: /* alias: %empty */ +#line 974 "yacc_sql.y" + { + (yyval.string) = nullptr; + } +#line 2902 "yacc_sql.cpp" + break; + + case 119: /* alias: ID */ +#line 977 "yacc_sql.y" + { + (yyval.string) = (yyvsp[0].string); + } +#line 2910 "yacc_sql.cpp" + break; + + case 120: /* alias: AS ID */ +#line 980 "yacc_sql.y" + { + (yyval.string) = (yyvsp[0].string); + } +#line 2918 "yacc_sql.cpp" + break; + + case 121: /* func_expr: ID LBRACE expression_list RBRACE */ +#line 986 "yacc_sql.y" + { + (yyval.expression) = new UnboundFunctionExpr((yyvsp[-3].string), std::move(*(yyvsp[-1].expression_list))); + (yyval.expression)->set_name(token_name(sql_string, &(yyloc))); + } +#line 2927 "yacc_sql.cpp" + break; + + case 122: /* sub_query_expr: LBRACE select_stmt RBRACE */ +#line 994 "yacc_sql.y" + { + (yyval.expression) = new SubQueryExpr((yyvsp[-1].sql_node)->selection); + } +#line 2935 "yacc_sql.cpp" + break; + + case 123: /* rel_attr: ID */ +#line 1000 "yacc_sql.y" + { + (yyval.rel_attr) = new RelAttrSqlNode; (yyval.rel_attr)->attribute_name = (yyvsp[0].string); free((yyvsp[0].string)); } -#line 2201 "yacc_sql.cpp" +#line 2945 "yacc_sql.cpp" break; - case 68: /* rel_attr: ID DOT ID */ -#line 544 "yacc_sql.y" - { - (yyval.rel_attr) = new RelAttrSqlNode; + case 124: /* rel_attr: ID DOT ID */ +#line 1005 "yacc_sql.y" + { + (yyval.rel_attr) = new RelAttrSqlNode; (yyval.rel_attr)->relation_name = (yyvsp[-2].string); (yyval.rel_attr)->attribute_name = (yyvsp[0].string); free((yyvsp[-2].string)); free((yyvsp[0].string)); } -#line 2213 "yacc_sql.cpp" +#line 2957 "yacc_sql.cpp" break; - case 69: /* relation: ID */ -#line 554 "yacc_sql.y" - { + case 125: /* relation: ID */ +#line 1015 "yacc_sql.y" + { (yyval.string) = (yyvsp[0].string); } -#line 2221 "yacc_sql.cpp" +#line 2965 "yacc_sql.cpp" break; - case 70: /* rel_list: relation */ -#line 559 "yacc_sql.y" - { - (yyval.relation_list) = new std::vector(); - (yyval.relation_list)->push_back((yyvsp[0].string)); - free((yyvsp[0].string)); + case 126: /* rel_list: relation alias */ +#line 1021 "yacc_sql.y" + { + (yyval.relation_list) = new std::vector(); + if (nullptr != (yyvsp[0].string)) { + (yyval.relation_list)->emplace_back((yyvsp[-1].string), (yyvsp[0].string)); + free((yyvsp[0].string)); + } else { + (yyval.relation_list)->emplace_back((yyvsp[-1].string)); + } + free((yyvsp[-1].string)); } -#line 2231 "yacc_sql.cpp" +#line 2980 "yacc_sql.cpp" break; - case 71: /* rel_list: relation COMMA rel_list */ -#line 564 "yacc_sql.y" - { + case 127: /* rel_list: relation alias COMMA rel_list */ +#line 1031 "yacc_sql.y" + { if ((yyvsp[0].relation_list) != nullptr) { (yyval.relation_list) = (yyvsp[0].relation_list); } else { - (yyval.relation_list) = new std::vector; + (yyval.relation_list) = new std::vector; } + if (nullptr != (yyvsp[-2].string)) { + (yyval.relation_list) + ->insert((yyval.relation_list)->begin(), RelationNode((yyvsp[-3].string), (yyvsp[-2].string))); + free((yyvsp[-2].string)); + } else { + (yyval.relation_list)->insert((yyval.relation_list)->begin(), RelationNode((yyvsp[-3].string))); + } + free((yyvsp[-3].string)); + } +#line 2999 "yacc_sql.cpp" + break; - (yyval.relation_list)->insert((yyval.relation_list)->begin(), (yyvsp[-2].string)); + case 128: /* join_clauses: relation ON condition */ +#line 1049 "yacc_sql.y" + { + (yyval.join_clauses) = new JoinSqlNode; + (yyval.join_clauses)->relations.emplace_back((yyvsp[-2].string)); + (yyval.join_clauses)->conditions = std::unique_ptr((yyvsp[0].expression)); free((yyvsp[-2].string)); } -#line 2246 "yacc_sql.cpp" +#line 3010 "yacc_sql.cpp" break; - case 72: /* where: %empty */ -#line 578 "yacc_sql.y" + case 129: /* join_clauses: relation ON condition INNER JOIN join_clauses */ +#line 1056 "yacc_sql.y" { - (yyval.condition_list) = nullptr; + (yyval.join_clauses) = (yyvsp[0].join_clauses); + (yyval.join_clauses)->relations.emplace_back((yyvsp[-5].string)); + auto ptr = (yyval.join_clauses)->conditions.release(); + (yyval.join_clauses)->conditions = + std::make_unique(ConjunctionExpr::Type::AND, ptr, (yyvsp[-3].expression)); + free((yyvsp[-5].string)); } -#line 2254 "yacc_sql.cpp" +#line 3022 "yacc_sql.cpp" break; - case 73: /* where: WHERE condition_list */ -#line 581 "yacc_sql.y" - { - (yyval.condition_list) = (yyvsp[0].condition_list); + case 130: /* where: %empty */ +#line 1067 "yacc_sql.y" + { + (yyval.expression) = nullptr; } -#line 2262 "yacc_sql.cpp" +#line 3030 "yacc_sql.cpp" break; - case 74: /* condition_list: %empty */ -#line 587 "yacc_sql.y" + case 131: /* where: WHERE condition */ +#line 1070 "yacc_sql.y" { - (yyval.condition_list) = nullptr; + (yyval.expression) = (yyvsp[0].expression); } -#line 2270 "yacc_sql.cpp" +#line 3038 "yacc_sql.cpp" break; - case 75: /* condition_list: condition */ -#line 590 "yacc_sql.y" - { - (yyval.condition_list) = new std::vector; - (yyval.condition_list)->emplace_back(*(yyvsp[0].condition)); - delete (yyvsp[0].condition); + case 132: /* condition: expression comp_op expression */ +#line 1077 "yacc_sql.y" + { + (yyval.expression) = new ComparisonExpr((yyvsp[-1].comp), (yyvsp[-2].expression), (yyvsp[0].expression)); } -#line 2280 "yacc_sql.cpp" +#line 3046 "yacc_sql.cpp" break; - case 76: /* condition_list: condition AND condition_list */ -#line 595 "yacc_sql.y" - { - (yyval.condition_list) = (yyvsp[0].condition_list); - (yyval.condition_list)->emplace_back(*(yyvsp[-2].condition)); - delete (yyvsp[-2].condition); + case 133: /* condition: comp_op expression */ +#line 1081 "yacc_sql.y" + { + Value val; + val.set_null(true); + ValueExpr *temp_expr = new ValueExpr(val); + (yyval.expression) = new ComparisonExpr((yyvsp[-1].comp), temp_expr, (yyvsp[0].expression)); } -#line 2290 "yacc_sql.cpp" +#line 3057 "yacc_sql.cpp" break; - case 77: /* condition: rel_attr comp_op value */ -#line 603 "yacc_sql.y" + case 134: /* condition: condition AND condition */ +#line 1088 "yacc_sql.y" { - (yyval.condition) = new ConditionSqlNode; - (yyval.condition)->left_is_attr = 1; - (yyval.condition)->left_attr = *(yyvsp[-2].rel_attr); - (yyval.condition)->right_is_attr = 0; - (yyval.condition)->right_value = *(yyvsp[0].value); - (yyval.condition)->comp = (yyvsp[-1].comp); + (yyval.expression) = + new ConjunctionExpr(ConjunctionExpr::Type::AND, (yyvsp[-2].expression), (yyvsp[0].expression)); + } +#line 3065 "yacc_sql.cpp" + break; - delete (yyvsp[-2].rel_attr); - delete (yyvsp[0].value); + case 135: /* condition: condition OR condition */ +#line 1092 "yacc_sql.y" + { + (yyval.expression) = + new ConjunctionExpr(ConjunctionExpr::Type::OR, (yyvsp[-2].expression), (yyvsp[0].expression)); } -#line 2306 "yacc_sql.cpp" +#line 3073 "yacc_sql.cpp" break; - case 78: /* condition: value comp_op value */ -#line 615 "yacc_sql.y" + case 136: /* comp_op: EQ */ +#line 1098 "yacc_sql.y" { - (yyval.condition) = new ConditionSqlNode; - (yyval.condition)->left_is_attr = 0; - (yyval.condition)->left_value = *(yyvsp[-2].value); - (yyval.condition)->right_is_attr = 0; - (yyval.condition)->right_value = *(yyvsp[0].value); - (yyval.condition)->comp = (yyvsp[-1].comp); + (yyval.comp) = EQUAL_TO; + } +#line 3079 "yacc_sql.cpp" + break; - delete (yyvsp[-2].value); - delete (yyvsp[0].value); + case 137: /* comp_op: LT */ +#line 1099 "yacc_sql.y" + { + (yyval.comp) = LESS_THAN; } -#line 2322 "yacc_sql.cpp" +#line 3085 "yacc_sql.cpp" break; - case 79: /* condition: rel_attr comp_op rel_attr */ -#line 627 "yacc_sql.y" + case 138: /* comp_op: GT */ +#line 1100 "yacc_sql.y" { - (yyval.condition) = new ConditionSqlNode; - (yyval.condition)->left_is_attr = 1; - (yyval.condition)->left_attr = *(yyvsp[-2].rel_attr); - (yyval.condition)->right_is_attr = 1; - (yyval.condition)->right_attr = *(yyvsp[0].rel_attr); - (yyval.condition)->comp = (yyvsp[-1].comp); + (yyval.comp) = GREAT_THAN; + } +#line 3091 "yacc_sql.cpp" + break; - delete (yyvsp[-2].rel_attr); - delete (yyvsp[0].rel_attr); + case 139: /* comp_op: LE */ +#line 1101 "yacc_sql.y" + { + (yyval.comp) = LESS_EQUAL; } -#line 2338 "yacc_sql.cpp" +#line 3097 "yacc_sql.cpp" break; - case 80: /* condition: value comp_op rel_attr */ -#line 639 "yacc_sql.y" + case 140: /* comp_op: GE */ +#line 1102 "yacc_sql.y" { - (yyval.condition) = new ConditionSqlNode; - (yyval.condition)->left_is_attr = 0; - (yyval.condition)->left_value = *(yyvsp[-2].value); - (yyval.condition)->right_is_attr = 1; - (yyval.condition)->right_attr = *(yyvsp[0].rel_attr); - (yyval.condition)->comp = (yyvsp[-1].comp); + (yyval.comp) = GREAT_EQUAL; + } +#line 3103 "yacc_sql.cpp" + break; - delete (yyvsp[-2].value); - delete (yyvsp[0].rel_attr); + case 141: /* comp_op: NE */ +#line 1103 "yacc_sql.y" + { + (yyval.comp) = NOT_EQUAL; } -#line 2354 "yacc_sql.cpp" +#line 3109 "yacc_sql.cpp" break; - case 81: /* comp_op: EQ */ -#line 653 "yacc_sql.y" - { (yyval.comp) = EQUAL_TO; } -#line 2360 "yacc_sql.cpp" + case 142: /* comp_op: IS */ +#line 1104 "yacc_sql.y" + { + (yyval.comp) = IS_OP; + } +#line 3115 "yacc_sql.cpp" break; - case 82: /* comp_op: LT */ -#line 654 "yacc_sql.y" - { (yyval.comp) = LESS_THAN; } -#line 2366 "yacc_sql.cpp" + case 143: /* comp_op: IS NOT */ +#line 1105 "yacc_sql.y" + { + (yyval.comp) = IS_NOT_OP; + } +#line 3121 "yacc_sql.cpp" break; - case 83: /* comp_op: GT */ -#line 655 "yacc_sql.y" - { (yyval.comp) = GREAT_THAN; } -#line 2372 "yacc_sql.cpp" + case 144: /* comp_op: LIKE */ +#line 1106 "yacc_sql.y" + { + (yyval.comp) = LIKE_OP; + } +#line 3127 "yacc_sql.cpp" break; - case 84: /* comp_op: LE */ -#line 656 "yacc_sql.y" - { (yyval.comp) = LESS_EQUAL; } -#line 2378 "yacc_sql.cpp" + case 145: /* comp_op: NOT LIKE */ +#line 1107 "yacc_sql.y" + { + (yyval.comp) = NOT_LIKE_OP; + } +#line 3133 "yacc_sql.cpp" break; - case 85: /* comp_op: GE */ -#line 657 "yacc_sql.y" - { (yyval.comp) = GREAT_EQUAL; } -#line 2384 "yacc_sql.cpp" + case 146: /* comp_op: IN */ +#line 1108 "yacc_sql.y" + { + (yyval.comp) = IN_OP; + } +#line 3139 "yacc_sql.cpp" break; - case 86: /* comp_op: NE */ -#line 658 "yacc_sql.y" - { (yyval.comp) = NOT_EQUAL; } -#line 2390 "yacc_sql.cpp" + case 147: /* comp_op: NOT IN */ +#line 1109 "yacc_sql.y" + { + (yyval.comp) = NOT_IN_OP; + } +#line 3145 "yacc_sql.cpp" break; - case 87: /* group_by: %empty */ -#line 664 "yacc_sql.y" + case 148: /* comp_op: EXISTS */ +#line 1110 "yacc_sql.y" + { + (yyval.comp) = EXISTS_OP; + } +#line 3151 "yacc_sql.cpp" + break; + + case 149: /* comp_op: NOT EXISTS */ +#line 1111 "yacc_sql.y" + { + (yyval.comp) = NOT_EXISTS_OP; + } +#line 3157 "yacc_sql.cpp" + break; + + case 150: /* opt_order_by: %empty */ +#line 1116 "yacc_sql.y" + { + (yyval.orderby_list) = nullptr; + } +#line 3165 "yacc_sql.cpp" + break; + + case 151: /* opt_order_by: ORDER BY sort_list */ +#line 1120 "yacc_sql.y" + { + (yyval.orderby_list) = (yyvsp[0].orderby_list); + std::reverse((yyval.orderby_list)->begin(), (yyval.orderby_list)->end()); + } +#line 3174 "yacc_sql.cpp" + break; + + case 152: /* sort_list: sort_unit */ +#line 1128 "yacc_sql.y" + { + (yyval.orderby_list) = new std::vector; + (yyval.orderby_list)->emplace_back(std::move(*(yyvsp[0].orderby_unit))); + } +#line 3183 "yacc_sql.cpp" + break; + + case 153: /* sort_list: sort_unit COMMA sort_list */ +#line 1133 "yacc_sql.y" + { + (yyvsp[0].orderby_list)->emplace_back(std::move(*(yyvsp[-2].orderby_unit))); + (yyval.orderby_list) = (yyvsp[0].orderby_list); + } +#line 3192 "yacc_sql.cpp" + break; + + case 154: /* sort_unit: expression */ +#line 1141 "yacc_sql.y" + { + (yyval.orderby_unit) = new OrderBySqlNode(); + (yyval.orderby_unit)->expr = std::unique_ptr((yyvsp[0].expression)); + (yyval.orderby_unit)->is_asc = true; + } +#line 3202 "yacc_sql.cpp" + break; + + case 155: /* sort_unit: expression DESC */ +#line 1147 "yacc_sql.y" + { + (yyval.orderby_unit) = new OrderBySqlNode(); + (yyval.orderby_unit)->expr = std::unique_ptr((yyvsp[-1].expression)); + (yyval.orderby_unit)->is_asc = false; + } +#line 3212 "yacc_sql.cpp" + break; + + case 156: /* sort_unit: expression ASC */ +#line 1153 "yacc_sql.y" + { + (yyval.orderby_unit) = new OrderBySqlNode(); // 默认是升序 + (yyval.orderby_unit)->expr = std::unique_ptr((yyvsp[-1].expression)); + (yyval.orderby_unit)->is_asc = true; + } +#line 3222 "yacc_sql.cpp" + break; + + case 157: /* group_by: %empty */ +#line 1162 "yacc_sql.y" { (yyval.expression_list) = nullptr; } -#line 2398 "yacc_sql.cpp" +#line 3230 "yacc_sql.cpp" break; - case 88: /* load_data_stmt: LOAD DATA INFILE SSS INTO TABLE ID */ -#line 670 "yacc_sql.y" + case 158: /* group_by: GROUP BY expression_list */ +#line 1166 "yacc_sql.y" { - char *tmp_file_name = common::substr((yyvsp[-3].string), 1, strlen((yyvsp[-3].string)) - 2); - - (yyval.sql_node) = new ParsedSqlNode(SCF_LOAD_DATA); - (yyval.sql_node)->load_data.relation_name = (yyvsp[0].string); - (yyval.sql_node)->load_data.file_name = tmp_file_name; - free((yyvsp[0].string)); - free(tmp_file_name); + (yyval.expression_list) = (yyvsp[0].expression_list); + } +#line 3238 "yacc_sql.cpp" + break; + + case 159: /* opt_having: %empty */ +#line 1173 "yacc_sql.y" + { + (yyval.expression) = nullptr; + } +#line 3246 "yacc_sql.cpp" + break; + + case 160: /* opt_having: HAVING condition */ +#line 1177 "yacc_sql.y" + { + (yyval.expression) = (yyvsp[0].expression); } -#line 2412 "yacc_sql.cpp" +#line 3254 "yacc_sql.cpp" break; - case 89: /* explain_stmt: EXPLAIN command_wrapper */ -#line 683 "yacc_sql.y" + case 161: /* opt_limit: %empty */ +#line 1184 "yacc_sql.y" { - (yyval.sql_node) = new ParsedSqlNode(SCF_EXPLAIN); + (yyval.limited_info) = nullptr; + } +#line 3262 "yacc_sql.cpp" + break; + + case 162: /* opt_limit: LIMIT NUMBER */ +#line 1188 "yacc_sql.y" + { + (yyval.limited_info) = new LimitSqlNode(); + (yyval.limited_info)->number = (yyvsp[0].number); + } +#line 3271 "yacc_sql.cpp" + break; + + case 163: /* explain_stmt: EXPLAIN command_wrapper */ +#line 1196 "yacc_sql.y" + { + (yyval.sql_node) = new ParsedSqlNode(SCF_EXPLAIN); (yyval.sql_node)->explain.sql_node = std::unique_ptr((yyvsp[0].sql_node)); } -#line 2421 "yacc_sql.cpp" +#line 3280 "yacc_sql.cpp" break; - case 90: /* set_variable_stmt: SET ID EQ value */ -#line 691 "yacc_sql.y" + case 164: /* set_variable_stmt: SET ID EQ value */ +#line 1204 "yacc_sql.y" { - (yyval.sql_node) = new ParsedSqlNode(SCF_SET_VARIABLE); + (yyval.sql_node) = new ParsedSqlNode(SCF_SET_VARIABLE); (yyval.sql_node)->set_variable.name = (yyvsp[-2].string); (yyval.sql_node)->set_variable.value = *(yyvsp[0].value); free((yyvsp[-2].string)); delete (yyvsp[0].value); } -#line 2433 "yacc_sql.cpp" +#line 3292 "yacc_sql.cpp" break; +#line 3296 "yacc_sql.cpp" -#line 2437 "yacc_sql.cpp" - - default: break; - } + default: break; + } /* User semantic actions sometimes alter yychar, and that requires that yytoken be updated with the new translation. We take the approach of translating immediately before every use of yytoken. @@ -2448,9 +5688,9 @@ YYLTYPE yylloc = yyloc_default; case of YYERROR or YYBACKUP, subsequent parser actions might lead to an incorrect destructor call or verbose syntax error message before the lookahead is translated. */ - YY_SYMBOL_PRINT ("-> $$ =", YY_CAST (yysymbol_kind_t, yyr1[yyn]), &yyval, &yyloc); + YY_SYMBOL_PRINT("-> $$ =", YY_CAST(yysymbol_kind_t, yyr1[yyn]), &yyval, &yyloc); - YYPOPSTACK (yylen); + YYPOPSTACK(yylen); yylen = 0; *++yyvsp = yyval; @@ -2461,84 +5701,67 @@ YYLTYPE yylloc = yyloc_default; number reduced by. */ { const int yylhs = yyr1[yyn] - YYNTOKENS; - const int yyi = yypgoto[yylhs] + *yyssp; - yystate = (0 <= yyi && yyi <= YYLAST && yycheck[yyi] == *yyssp - ? yytable[yyi] - : yydefgoto[yylhs]); + const int yyi = yypgoto[yylhs] + *yyssp; + yystate = (0 <= yyi && yyi <= YYLAST && yycheck[yyi] == *yyssp ? yytable[yyi] : yydefgoto[yylhs]); } goto yynewstate; - /*--------------------------------------. | yyerrlab -- here on detecting error. | `--------------------------------------*/ yyerrlab: /* Make sure we have latest lookahead translation. See comments at user semantic actions for why this is necessary. */ - yytoken = yychar == YYEMPTY ? YYSYMBOL_YYEMPTY : YYTRANSLATE (yychar); + yytoken = yychar == YYEMPTY ? YYSYMBOL_YYEMPTY : YYTRANSLATE(yychar); /* If not already recovering from an error, report this error. */ - if (!yyerrstatus) - { - ++yynerrs; - { - yypcontext_t yyctx - = {yyssp, yytoken, &yylloc}; - char const *yymsgp = YY_("syntax error"); - int yysyntax_error_status; - yysyntax_error_status = yysyntax_error (&yymsg_alloc, &yymsg, &yyctx); - if (yysyntax_error_status == 0) - yymsgp = yymsg; - else if (yysyntax_error_status == -1) - { - if (yymsg != yymsgbuf) - YYSTACK_FREE (yymsg); - yymsg = YY_CAST (char *, - YYSTACK_ALLOC (YY_CAST (YYSIZE_T, yymsg_alloc))); - if (yymsg) - { - yysyntax_error_status - = yysyntax_error (&yymsg_alloc, &yymsg, &yyctx); - yymsgp = yymsg; - } - else - { - yymsg = yymsgbuf; - yymsg_alloc = sizeof yymsgbuf; - yysyntax_error_status = YYENOMEM; - } - } - yyerror (&yylloc, sql_string, sql_result, scanner, yymsgp); - if (yysyntax_error_status == YYENOMEM) - YYNOMEM; + if (!yyerrstatus) { + ++yynerrs; + { + yypcontext_t yyctx = {yyssp, yytoken, &yylloc}; + char const *yymsgp = YY_("syntax error"); + int yysyntax_error_status; + yysyntax_error_status = yysyntax_error(&yymsg_alloc, &yymsg, &yyctx); + if (yysyntax_error_status == 0) + yymsgp = yymsg; + else if (yysyntax_error_status == -1) { + if (yymsg != yymsgbuf) + YYSTACK_FREE(yymsg); + yymsg = YY_CAST(char *, YYSTACK_ALLOC(YY_CAST(YYSIZE_T, yymsg_alloc))); + if (yymsg) { + yysyntax_error_status = yysyntax_error(&yymsg_alloc, &yymsg, &yyctx); + yymsgp = yymsg; + } else { + yymsg = yymsgbuf; + yymsg_alloc = sizeof yymsgbuf; + yysyntax_error_status = YYENOMEM; + } } + yyerror(&yylloc, sql_string, sql_result, scanner, yymsgp); + if (yysyntax_error_status == YYENOMEM) + YYNOMEM; } + } yyerror_range[1] = yylloc; - if (yyerrstatus == 3) - { - /* If just tried and failed to reuse lookahead token after an - error, discard it. */ + if (yyerrstatus == 3) { + /* If just tried and failed to reuse lookahead token after an + error, discard it. */ - if (yychar <= YYEOF) - { - /* Return failure if at end of input. */ - if (yychar == YYEOF) - YYABORT; - } - else - { - yydestruct ("Error: discarding", - yytoken, &yylval, &yylloc, sql_string, sql_result, scanner); - yychar = YYEMPTY; - } + if (yychar <= YYEOF) { + /* Return failure if at end of input. */ + if (yychar == YYEOF) + YYABORT; + } else { + yydestruct("Error: discarding", yytoken, &yylval, &yylloc, sql_string, sql_result, scanner); + yychar = YYEMPTY; } + } /* Else will try to reuse lookahead token after shifting the error token. */ goto yyerrlab1; - /*---------------------------------------------------. | yyerrorlab -- error raised explicitly by YYERROR. | `---------------------------------------------------*/ @@ -2551,45 +5774,40 @@ YYLTYPE yylloc = yyloc_default; /* Do not reclaim the symbols of the rule whose action triggered this YYERROR. */ - YYPOPSTACK (yylen); + YYPOPSTACK(yylen); yylen = 0; - YY_STACK_PRINT (yyss, yyssp); + YY_STACK_PRINT(yyss, yyssp); yystate = *yyssp; goto yyerrlab1; - /*-------------------------------------------------------------. | yyerrlab1 -- common code for both syntax error and YYERROR. | `-------------------------------------------------------------*/ yyerrlab1: - yyerrstatus = 3; /* Each real token shifted decrements this. */ + yyerrstatus = 3; /* Each real token shifted decrements this. */ /* Pop stack until we find a state that shifts the error token. */ - for (;;) - { - yyn = yypact[yystate]; - if (!yypact_value_is_default (yyn)) - { - yyn += YYSYMBOL_YYerror; - if (0 <= yyn && yyn <= YYLAST && yycheck[yyn] == YYSYMBOL_YYerror) - { - yyn = yytable[yyn]; - if (0 < yyn) - break; - } - } + for (;;) { + yyn = yypact[yystate]; + if (!yypact_value_is_default(yyn)) { + yyn += YYSYMBOL_YYerror; + if (0 <= yyn && yyn <= YYLAST && yycheck[yyn] == YYSYMBOL_YYerror) { + yyn = yytable[yyn]; + if (0 < yyn) + break; + } + } - /* Pop the current state because it cannot handle the error token. */ - if (yyssp == yyss) - YYABORT; + /* Pop the current state because it cannot handle the error token. */ + if (yyssp == yyss) + YYABORT; - yyerror_range[1] = *yylsp; - yydestruct ("Error: popping", - YY_ACCESSING_SYMBOL (yystate), yyvsp, yylsp, sql_string, sql_result, scanner); - YYPOPSTACK (1); - yystate = *yyssp; - YY_STACK_PRINT (yyss, yyssp); - } + yyerror_range[1] = *yylsp; + yydestruct("Error: popping", YY_ACCESSING_SYMBOL(yystate), yyvsp, yylsp, sql_string, sql_result, scanner); + YYPOPSTACK(1); + yystate = *yyssp; + YY_STACK_PRINT(yyss, yyssp); + } YY_IGNORE_MAYBE_UNINITIALIZED_BEGIN *++yyvsp = yylval; @@ -2597,15 +5815,14 @@ YYLTYPE yylloc = yyloc_default; yyerror_range[2] = yylloc; ++yylsp; - YYLLOC_DEFAULT (*yylsp, yyerror_range, 2); + YYLLOC_DEFAULT(*yylsp, yyerror_range, 2); /* Shift the error token. */ - YY_SYMBOL_PRINT ("Shifting", YY_ACCESSING_SYMBOL (yyn), yyvsp, yylsp); + YY_SYMBOL_PRINT("Shifting", YY_ACCESSING_SYMBOL(yyn), yyvsp, yylsp); yystate = yyn; goto yynewstate; - /*-------------------------------------. | yyacceptlab -- YYACCEPT comes here. | `-------------------------------------*/ @@ -2613,7 +5830,6 @@ YYLTYPE yylloc = yyloc_default; yyresult = 0; goto yyreturnlab; - /*-----------------------------------. | yyabortlab -- YYABORT comes here. | `-----------------------------------*/ @@ -2621,53 +5837,48 @@ YYLTYPE yylloc = yyloc_default; yyresult = 1; goto yyreturnlab; - /*-----------------------------------------------------------. | yyexhaustedlab -- YYNOMEM (memory exhaustion) comes here. | `-----------------------------------------------------------*/ yyexhaustedlab: - yyerror (&yylloc, sql_string, sql_result, scanner, YY_("memory exhausted")); + yyerror(&yylloc, sql_string, sql_result, scanner, YY_("memory exhausted")); yyresult = 2; goto yyreturnlab; - /*----------------------------------------------------------. | yyreturnlab -- parsing is finished, clean up and return. | `----------------------------------------------------------*/ yyreturnlab: - if (yychar != YYEMPTY) - { - /* Make sure we have latest lookahead translation. See comments at - user semantic actions for why this is necessary. */ - yytoken = YYTRANSLATE (yychar); - yydestruct ("Cleanup: discarding lookahead", - yytoken, &yylval, &yylloc, sql_string, sql_result, scanner); - } + if (yychar != YYEMPTY) { + /* Make sure we have latest lookahead translation. See comments at + user semantic actions for why this is necessary. */ + yytoken = YYTRANSLATE(yychar); + yydestruct("Cleanup: discarding lookahead", yytoken, &yylval, &yylloc, sql_string, sql_result, scanner); + } /* Do not reclaim the symbols of the rule whose action triggered this YYABORT or YYACCEPT. */ - YYPOPSTACK (yylen); - YY_STACK_PRINT (yyss, yyssp); - while (yyssp != yyss) - { - yydestruct ("Cleanup: popping", - YY_ACCESSING_SYMBOL (+*yyssp), yyvsp, yylsp, sql_string, sql_result, scanner); - YYPOPSTACK (1); - } + YYPOPSTACK(yylen); + YY_STACK_PRINT(yyss, yyssp); + while (yyssp != yyss) { + yydestruct("Cleanup: popping", YY_ACCESSING_SYMBOL(+*yyssp), yyvsp, yylsp, sql_string, sql_result, scanner); + YYPOPSTACK(1); + } #ifndef yyoverflow if (yyss != yyssa) - YYSTACK_FREE (yyss); + YYSTACK_FREE(yyss); #endif if (yymsg != yymsgbuf) - YYSTACK_FREE (yymsg); + YYSTACK_FREE(yymsg); return yyresult; } -#line 703 "yacc_sql.y" +#line 1216 "yacc_sql.y" //_____________________________________________________________________ extern void scan_string(const char *str, yyscan_t scanner); -int sql_parse(const char *s, ParsedSqlResult *sql_result) { +int sql_parse(const char *s, ParsedSqlResult *sql_result) +{ yyscan_t scanner; yylex_init(&scanner); scan_string(s, scanner); diff --git a/src/observer/sql/parser/yacc_sql.hpp b/src/observer/sql/parser/yacc_sql.hpp index bcc8eccf..58adfc3c 100644 --- a/src/observer/sql/parser/yacc_sql.hpp +++ b/src/observer/sql/parser/yacc_sql.hpp @@ -36,10 +36,10 @@ private implementation details that can be changed or removed. */ #ifndef YY_YY_YACC_SQL_HPP_INCLUDED -# define YY_YY_YACC_SQL_HPP_INCLUDED +#define YY_YY_YACC_SQL_HPP_INCLUDED /* Debug traces. */ #ifndef YYDEBUG -# define YYDEBUG 0 +#define YYDEBUG 0 #endif #if YYDEBUG extern int yydebug; @@ -47,103 +47,142 @@ extern int yydebug; /* Token kinds. */ #ifndef YYTOKENTYPE -# define YYTOKENTYPE - enum yytokentype - { - YYEMPTY = -2, - YYEOF = 0, /* "end of file" */ - YYerror = 256, /* error */ - YYUNDEF = 257, /* "invalid token" */ - SEMICOLON = 258, /* SEMICOLON */ - BY = 259, /* BY */ - CREATE = 260, /* CREATE */ - DROP = 261, /* DROP */ - GROUP = 262, /* GROUP */ - TABLE = 263, /* TABLE */ - TABLES = 264, /* TABLES */ - INDEX = 265, /* INDEX */ - CALC = 266, /* CALC */ - SELECT = 267, /* SELECT */ - DESC = 268, /* DESC */ - SHOW = 269, /* SHOW */ - SYNC = 270, /* SYNC */ - INSERT = 271, /* INSERT */ - DELETE = 272, /* DELETE */ - UPDATE = 273, /* UPDATE */ - LBRACE = 274, /* LBRACE */ - RBRACE = 275, /* RBRACE */ - COMMA = 276, /* COMMA */ - TRX_BEGIN = 277, /* TRX_BEGIN */ - TRX_COMMIT = 278, /* TRX_COMMIT */ - TRX_ROLLBACK = 279, /* TRX_ROLLBACK */ - INT_T = 280, /* INT_T */ - STRING_T = 281, /* STRING_T */ - FLOAT_T = 282, /* FLOAT_T */ - VECTOR_T = 283, /* VECTOR_T */ - HELP = 284, /* HELP */ - EXIT = 285, /* EXIT */ - DOT = 286, /* DOT */ - INTO = 287, /* INTO */ - VALUES = 288, /* VALUES */ - FROM = 289, /* FROM */ - WHERE = 290, /* WHERE */ - AND = 291, /* AND */ - SET = 292, /* SET */ - ON = 293, /* ON */ - LOAD = 294, /* LOAD */ - DATA = 295, /* DATA */ - INFILE = 296, /* INFILE */ - EXPLAIN = 297, /* EXPLAIN */ - STORAGE = 298, /* STORAGE */ - FORMAT = 299, /* FORMAT */ - EQ = 300, /* EQ */ - LT = 301, /* LT */ - GT = 302, /* GT */ - LE = 303, /* LE */ - GE = 304, /* GE */ - NE = 305, /* NE */ - NUMBER = 306, /* NUMBER */ - FLOAT = 307, /* FLOAT */ - ID = 308, /* ID */ - SSS = 309, /* SSS */ - UMINUS = 310 /* UMINUS */ - }; - typedef enum yytokentype yytoken_kind_t; +#define YYTOKENTYPE +enum yytokentype +{ + YYEMPTY = -2, + YYEOF = 0, /* "end of file" */ + YYerror = 256, /* error */ + YYUNDEF = 257, /* "invalid token" */ + SEMICOLON = 258, /* SEMICOLON */ + AS = 259, /* AS */ + ASC = 260, /* ASC */ + BY = 261, /* BY */ + CREATE = 262, /* CREATE */ + DROP = 263, /* DROP */ + EXISTS = 264, /* EXISTS */ + GROUP = 265, /* GROUP */ + HAVING = 266, /* HAVING */ + ORDER = 267, /* ORDER */ + TABLE = 268, /* TABLE */ + TABLES = 269, /* TABLES */ + INDEX = 270, /* INDEX */ + CALC = 271, /* CALC */ + SELECT = 272, /* SELECT */ + DESC = 273, /* DESC */ + SHOW = 274, /* SHOW */ + SYNC = 275, /* SYNC */ + INSERT = 276, /* INSERT */ + DELETE = 277, /* DELETE */ + UPDATE = 278, /* UPDATE */ + LBRACE = 279, /* LBRACE */ + RBRACE = 280, /* RBRACE */ + LSBRACE = 281, /* LSBRACE */ + RSBRACE = 282, /* RSBRACE */ + COMMA = 283, /* COMMA */ + TRX_BEGIN = 284, /* TRX_BEGIN */ + TRX_COMMIT = 285, /* TRX_COMMIT */ + TRX_ROLLBACK = 286, /* TRX_ROLLBACK */ + INT_T = 287, /* INT_T */ + IN = 288, /* IN */ + TRUE = 289, /* TRUE */ + FALSE = 290, /* FALSE */ + STRING_T = 291, /* STRING_T */ + FLOAT_T = 292, /* FLOAT_T */ + DATE_T = 293, /* DATE_T */ + TEXT_T = 294, /* TEXT_T */ + VECTOR_T = 295, /* VECTOR_T */ + NOT = 296, /* NOT */ + UNIQUE = 297, /* UNIQUE */ + NULL_T = 298, /* NULL_T */ + LIMIT = 299, /* LIMIT */ + NULLABLE = 300, /* NULLABLE */ + HELP = 301, /* HELP */ + QUOTE = 302, /* QUOTE */ + EXIT = 303, /* EXIT */ + DOT = 304, /* DOT */ + INTO = 305, /* INTO */ + VALUES = 306, /* VALUES */ + FROM = 307, /* FROM */ + WHERE = 308, /* WHERE */ + AND = 309, /* AND */ + OR = 310, /* OR */ + SET = 311, /* SET */ + ON = 312, /* ON */ + INFILE = 313, /* INFILE */ + EXPLAIN = 314, /* EXPLAIN */ + STORAGE = 315, /* STORAGE */ + FORMAT = 316, /* FORMAT */ + INNER = 317, /* INNER */ + JOIN = 318, /* JOIN */ + VIEW = 319, /* VIEW */ + WITH = 320, /* WITH */ + DISTANCE = 321, /* DISTANCE */ + TYPE = 322, /* TYPE */ + LISTS = 323, /* LISTS */ + PROBES = 324, /* PROBES */ + IVFFLAT = 325, /* IVFFLAT */ + EQ = 326, /* EQ */ + LT = 327, /* LT */ + GT = 328, /* GT */ + LE = 329, /* LE */ + GE = 330, /* GE */ + NE = 331, /* NE */ + LIKE = 332, /* LIKE */ + IS = 333, /* IS */ + NUMBER = 334, /* NUMBER */ + FLOAT = 335, /* FLOAT */ + ID = 336, /* ID */ + SSS = 337, /* SSS */ + UMINUS = 338 /* UMINUS */ +}; +typedef enum yytokentype yytoken_kind_t; #endif /* Value type. */ -#if ! defined YYSTYPE && ! defined YYSTYPE_IS_DECLARED +#if !defined YYSTYPE && !defined YYSTYPE_IS_DECLARED union YYSTYPE { -#line 117 "yacc_sql.y" - - ParsedSqlNode * sql_node; - ConditionSqlNode * condition; - Value * value; - enum CompOp comp; - RelAttrSqlNode * rel_attr; - std::vector * attr_infos; - AttrInfoSqlNode * attr_info; - Expression * expression; - std::vector> * expression_list; - std::vector * value_list; - std::vector * condition_list; - std::vector * rel_attr_list; - std::vector * relation_list; - char * string; - int number; - float floats; - -#line 138 "yacc_sql.hpp" - +#line 175 "yacc_sql.y" + + ParsedSqlNode *sql_node; + Value *value; + enum CompOp comp; + RelAttrSqlNode *rel_attr; + std::vector *attr_infos; + AttrInfoSqlNode *attr_info; + Expression *expression; + std::vector> *expression_list; + std::vector *value_list; + std::vector> *values_list; + SetClauseSqlNode *set_clause; + std::vector *set_clauses; + JoinSqlNode *join_clauses; + std::vector *rel_attr_list; + std::vector *relation_list; + OrderBySqlNode *orderby_unit; + std::vector *orderby_list; + LimitSqlNode *limited_info; + char *string; + int number; + float floats; + bool nullable_info; + std::vector *index_attr_list; + bool unique; + enum IndexType index_type; + VectorIndexConfig *vector_index_config; + float digits; + std::vector *digits_list; + +#line 178 "yacc_sql.hpp" }; typedef union YYSTYPE YYSTYPE; -# define YYSTYPE_IS_TRIVIAL 1 -# define YYSTYPE_IS_DECLARED 1 +#define YYSTYPE_IS_TRIVIAL 1 +#define YYSTYPE_IS_DECLARED 1 #endif /* Location type. */ -#if ! defined YYLTYPE && ! defined YYLTYPE_IS_DECLARED +#if !defined YYLTYPE && !defined YYLTYPE_IS_DECLARED typedef struct YYLTYPE YYLTYPE; struct YYLTYPE { @@ -152,14 +191,10 @@ struct YYLTYPE int last_line; int last_column; }; -# define YYLTYPE_IS_DECLARED 1 -# define YYLTYPE_IS_TRIVIAL 1 +#define YYLTYPE_IS_DECLARED 1 +#define YYLTYPE_IS_TRIVIAL 1 #endif - - - -int yyparse (const char * sql_string, ParsedSqlResult * sql_result, void * scanner); - +int yyparse(const char *sql_string, ParsedSqlResult *sql_result, void *scanner); #endif /* !YY_YY_YACC_SQL_HPP_INCLUDED */ diff --git a/src/observer/sql/parser/yacc_sql.y b/src/observer/sql/parser/yacc_sql.y index 9e392b62..47ef43f5 100644 --- a/src/observer/sql/parser/yacc_sql.y +++ b/src/observer/sql/parser/yacc_sql.y @@ -41,16 +41,46 @@ ArithmeticExpr *create_arithmetic_expression(ArithmeticExpr::Type type, return expr; } -UnboundAggregateExpr *create_aggregate_expression(const char *aggregate_name, - Expression *child, - const char *sql_string, - YYLTYPE *llocp) +UnboundFunctionExpr *create_aggregate_expression(const char *function_name, + std::vector> child, + const char *sql_string, + YYLTYPE *llocp) { - UnboundAggregateExpr *expr = new UnboundAggregateExpr(aggregate_name, child); + UnboundFunctionExpr *expr = new UnboundFunctionExpr(function_name, std::move(child)); expr->set_name(token_name(sql_string, llocp)); return expr; } +ParsedSqlNode *create_table_sql_node(char *table_name, + AttrInfoSqlNode* attr_def, + std::vector *attrinfos, + char* storage_format, + ParsedSqlNode *create_table_select) +{ + ParsedSqlNode *parsed_sql_node = new ParsedSqlNode(SCF_CREATE_TABLE); + CreateTableSqlNode &create_table = parsed_sql_node->create_table; + create_table.relation_name = table_name; + + if (attrinfos) { + create_table.attr_infos.swap(*attrinfos); + delete attrinfos; + } + if (attr_def) { + create_table.attr_infos.emplace_back(*attr_def); + std::reverse(create_table.attr_infos.begin(), create_table.attr_infos.end()); + delete attr_def; + } + if (storage_format != nullptr) { + create_table.storage_format = storage_format; + free(storage_format); + } + + if (create_table_select) { + create_table.create_table_select = std::make_unique(std::move(create_table_select->selection)); + } + + return parsed_sql_node; +} %} %define api.pure full @@ -65,10 +95,15 @@ UnboundAggregateExpr *create_aggregate_expression(const char *aggregate_name, //标识tokens %token SEMICOLON + AS + ASC BY CREATE DROP + EXISTS GROUP + HAVING + ORDER TABLE TABLES INDEX @@ -82,41 +117,63 @@ UnboundAggregateExpr *create_aggregate_expression(const char *aggregate_name, UPDATE LBRACE RBRACE + LSBRACE + RSBRACE COMMA TRX_BEGIN TRX_COMMIT TRX_ROLLBACK INT_T + IN + TRUE + FALSE STRING_T FLOAT_T + DATE_T + TEXT_T VECTOR_T + NOT + UNIQUE + NULL_T + LIMIT + NULLABLE HELP + QUOTE EXIT - DOT //QUOTE + DOT INTO VALUES FROM WHERE AND + OR SET ON - LOAD - DATA INFILE EXPLAIN STORAGE FORMAT + INNER + JOIN + VIEW + WITH + DISTANCE + TYPE + LISTS + PROBES + IVFFLAT EQ LT GT LE GE NE + LIKE + IS /** union 中定义各种数据类型，真实生成的代码也是union类型，所以不能有非POD类型的数据 **/ %union { ParsedSqlNode * sql_node; - ConditionSqlNode * condition; Value * value; enum CompOp comp; RelAttrSqlNode * rel_attr; @@ -125,12 +182,25 @@ UnboundAggregateExpr *create_aggregate_expression(const char *aggregate_name, Expression * expression; std::vector> * expression_list; std::vector * value_list; - std::vector * condition_list; + std::vector> * values_list; + SetClauseSqlNode * set_clause; + std::vector * set_clauses; + JoinSqlNode * join_clauses; std::vector * rel_attr_list; - std::vector * relation_list; + std::vector * relation_list; + OrderBySqlNode * orderby_unit; + std::vector * orderby_list; + LimitSqlNode * limited_info; char * string; int number; float floats; + bool nullable_info; + std::vector * index_attr_list; + bool unique; + enum IndexType index_type; + VectorIndexConfig * vector_index_config; + float digits; + std::vector * digits_list; } %token NUMBER @@ -141,22 +211,40 @@ UnboundAggregateExpr *create_aggregate_expression(const char *aggregate_name, /** type 定义了各种解析后的结果输出的是什么类型。类型对应了 union 中的定义的成员变量名称 **/ %type type -%type condition +%type digits +%type digits_list %type value -%type number +%type nonnegative_value %type relation +%type alias %type comp_op %type rel_attr +%type nullable_constraint %type attr_def_list %type attr_def %type value_list -%type where -%type condition_list +%type values_list +%type condition %type storage_format %type rel_list %type expression +%type where +%type func_expr +%type sub_query_expr %type expression_list %type group_by +%type opt_having +%type set_clause +%type set_clauses +%type join_clauses +%type sort_unit +%type sort_list +%type opt_order_by +%type opt_limit +%type attr_list +%type opt_unique +%type index_type +%type vector_index_config %type calc_stmt %type select_stmt %type insert_stmt @@ -168,11 +256,13 @@ UnboundAggregateExpr *create_aggregate_expression(const char *aggregate_name, %type desc_table_stmt %type create_index_stmt %type drop_index_stmt +%type show_index_stmt +%type create_view_stmt +%type drop_view_stmt %type sync_stmt %type begin_stmt %type commit_stmt %type rollback_stmt -%type load_data_stmt %type explain_stmt %type set_variable_stmt %type help_stmt @@ -181,6 +271,9 @@ UnboundAggregateExpr *create_aggregate_expression(const char *aggregate_name, // commands should be a list but I use a single command instead %type commands +%left OR +%left AND +%left EQ LT GT LE GE NE %left '+' '-' %left '*' '/' %nonassoc UMINUS @@ -194,8 +287,8 @@ commands: command_wrapper opt_semicolon //commands or sqls. parser starts here. ; command_wrapper: - calc_stmt - | select_stmt + select_stmt + | calc_stmt | insert_stmt | update_stmt | delete_stmt @@ -205,11 +298,13 @@ command_wrapper: | desc_table_stmt | create_index_stmt | drop_index_stmt + | show_index_stmt + | create_view_stmt + | drop_view_stmt | sync_stmt | begin_stmt | commit_stmt | rollback_stmt - | load_data_stmt | explain_stmt | set_variable_stmt | help_stmt @@ -272,17 +367,96 @@ desc_table_stmt: } ; -create_index_stmt: /*create index 语句的语法解析树*/ - CREATE INDEX ID ON ID LBRACE ID RBRACE +show_index_stmt: + SHOW INDEX FROM relation + { + $$ = new ParsedSqlNode(SCF_SHOW_INDEX); + ShowIndexSqlNode &show_index = $$->show_index; + show_index.relation_name = $4; + free($4); + } + ; + +create_index_stmt: + CREATE opt_unique INDEX ID ON ID LBRACE attr_list RBRACE { $$ = new ParsedSqlNode(SCF_CREATE_INDEX); CreateIndexSqlNode &create_index = $$->create_index; - create_index.index_name = $3; - create_index.relation_name = $5; - create_index.attribute_name = $7; - free($3); - free($5); - free($7); + create_index.unique = $2; // 用 opt_unique 的返回值来确定是否 UNIQUE + create_index.index_name = $4; + create_index.relation_name = $6; + create_index.attribute_name.swap(*$8); // $8 是 vector 类型 + delete $8; // 释放指针 + free($4); + free($6); + } + | CREATE VECTOR_T INDEX ID ON ID LBRACE attr_list RBRACE WITH vector_index_config + { + $$ = new ParsedSqlNode(SCF_CREATE_INDEX); + CreateIndexSqlNode &create_index = $$->create_index; + create_index.unique = false; // 向量索引不支持 + create_index.index_name = $4; + create_index.relation_name = $6; + create_index.attribute_name.swap(*$8); // $8 是 vector 类型 + create_index.vector_index_config = std::move(*$11); + delete $8; // 释放指针 + free($4); + free($6); + } + ; + +opt_unique: + UNIQUE { $$ = true; } + | /* 空 */ { $$ = false; } + ; + +index_type: + IVFFLAT + { + $$ = IndexType::VectorIVFFlatIndex; + } + ; + +vector_index_config: + LBRACE DISTANCE EQ ID COMMA TYPE EQ index_type RBRACE + { + $$ = new VectorIndexConfig; + $$->distance_fn = $4; + $$->index_type = $8; + free($4); + } + | LBRACE DISTANCE EQ ID COMMA TYPE EQ index_type COMMA LISTS EQ value COMMA PROBES EQ value RBRACE + { + $$ = new VectorIndexConfig; + $$->distance_fn = $4; + $$->index_type = $8; + $$->lists = std::move(*$12); + $$->probes = std::move(*$16); + free($4); + } + | LBRACE TYPE EQ index_type COMMA DISTANCE EQ ID COMMA LISTS EQ value COMMA PROBES EQ value RBRACE + { + $$ = new VectorIndexConfig; + $$->distance_fn = $8; + $$->index_type = $4; + $$->lists = std::move(*$12); + $$->probes = std::move(*$16); + free($8); + } + ; + +attr_list: + ID + { + $$ = new std::vector; // 创建一个新的 vector + $$->emplace_back($1); // 将列名加入 vector + free($1); + } + | ID COMMA attr_list + { + $$ = $3; // 使用现有的 vector + $$->emplace($$->begin(), $1); // 将新列名加入 vector 开头 + free($1); } ; @@ -297,28 +471,57 @@ drop_index_stmt: /*drop index 语句的语法解析树*/ } ; create_table_stmt: /*create table 语句的语法解析树*/ - CREATE TABLE ID LBRACE attr_def attr_def_list RBRACE storage_format + CREATE TABLE ID LBRACE attr_def attr_def_list RBRACE storage_format AS select_stmt { - $$ = new ParsedSqlNode(SCF_CREATE_TABLE); - CreateTableSqlNode &create_table = $$->create_table; - create_table.relation_name = $3; - free($3); + $$ = create_table_sql_node($3, $5, $6, $8, $10); + } + | CREATE TABLE ID LBRACE attr_def attr_def_list RBRACE storage_format select_stmt + { + $$ = create_table_sql_node($3, $5, $6, $8, $9); + } + | CREATE TABLE ID LBRACE attr_def attr_def_list RBRACE storage_format + { + $$ = create_table_sql_node($3, $5, $6, $8, nullptr); + } + | CREATE TABLE ID storage_format AS select_stmt + { + $$ = create_table_sql_node($3, nullptr, nullptr, $4, $6); + } + | CREATE TABLE ID storage_format select_stmt + { + $$ = create_table_sql_node($3, nullptr, nullptr, $4, $5); + } + ; - std::vector *src_attrs = $6; +create_view_stmt: + CREATE VIEW ID AS select_stmt + { + $$ = new ParsedSqlNode(SCF_CREATE_VIEW); + CreateViewSqlNode &create_view = $$->create_view; + create_view.relation_name = $3; + create_view.create_view_select = std::make_unique(std::move($5->selection)); + free($3); + } + | CREATE VIEW ID LBRACE attr_list RBRACE AS select_stmt + { + $$ = new ParsedSqlNode(SCF_CREATE_VIEW); + CreateViewSqlNode &create_view = $$->create_view; + create_view.relation_name = $3; + create_view.attribute_names = std::move(*$5); + create_view.create_view_select = std::make_unique(std::move($8->selection)); + free($3); + } + ; - if (src_attrs != nullptr) { - create_table.attr_infos.swap(*src_attrs); - delete src_attrs; - } - create_table.attr_infos.emplace_back(*$5); - std::reverse(create_table.attr_infos.begin(), create_table.attr_infos.end()); - delete $5; - if ($8 != nullptr) { - create_table.storage_format = $8; - free($8); - } +drop_view_stmt: + DROP VIEW ID + { + $$ = new ParsedSqlNode(SCF_DROP_VIEW); + $$->drop_view.relation_name = $3; + free($3); } ; + attr_def_list: /* empty */ { @@ -337,79 +540,213 @@ attr_def_list: ; attr_def: - ID type LBRACE number RBRACE + ID type LBRACE NUMBER RBRACE nullable_constraint { $$ = new AttrInfoSqlNode; - $$->type = (AttrType)$2; $$->name = $1; - $$->length = $4; + $$->type = (AttrType)$2; + if ($$->type == AttrType::CHARS) { + $$->length = $4; + } else if ($$->type == AttrType::VECTORS) { + $$->length = sizeof(float) * $4; + } else { + ASSERT(false, "$$->type is invalid."); + } + $$->nullable = $6; + if ($$->nullable) { + $$->length++; + } free($1); } - | ID type + | ID type nullable_constraint { $$ = new AttrInfoSqlNode; $$->type = (AttrType)$2; $$->name = $1; - $$->length = 4; + if ($$->type == AttrType::INTS) { + $$->length = sizeof(int); + } else if ($$->type == AttrType::FLOATS) { + $$->length = sizeof(float); + } else if ($$->type == AttrType::DATES) { + $$->length = sizeof(int); + } else if ($$->type == AttrType::CHARS) { + $$->length = sizeof(char); + } else if ($$->type == AttrType::VECTORS) { + $$->length = sizeof(float) * 1; + } else if ($$->type == AttrType::TEXTS) { + $$->length = 65535; + } else { + ASSERT(false, "$$->type is invalid."); + } + $$->nullable = $3; // 处理NULL/NOT NULL标记 + if ($$->nullable) { + $$->length++; + } free($1); } ; -number: - NUMBER {$$ = $1;} + +nullable_constraint: + NOT NULL_T + { + $$ = false; // NOT NULL 对应的可空性为 false + } + | NULLABLE + { + $$ = true; // NULLABLE 对应的可空性为 true 2022 + } + | NULL_T + { + $$ = true; // NULL 对应的可空性也为 true 2023 + } + | /* empty */ + { + $$ = true; // 默认情况为 NULL + } ; + type: - INT_T { $$ = static_cast(AttrType::INTS); } - | STRING_T { $$ = static_cast(AttrType::CHARS); } - | FLOAT_T { $$ = static_cast(AttrType::FLOATS); } - | VECTOR_T { $$ = static_cast(AttrType::VECTORS); } + INT_T { $$ = static_cast(AttrType::INTS); } + | STRING_T { $$ = static_cast(AttrType::CHARS); } + | FLOAT_T { $$ = static_cast(AttrType::FLOATS); } + | DATE_T { $$ = static_cast(AttrType::DATES); } + | TEXT_T { $$ = static_cast(AttrType::TEXTS); } + | VECTOR_T { $$ = static_cast(AttrType::VECTORS); } ; + insert_stmt: /*insert 语句的语法解析树*/ - INSERT INTO ID VALUES LBRACE value value_list RBRACE + INSERT INTO ID VALUES values_list { $$ = new ParsedSqlNode(SCF_INSERT); $$->insertion.relation_name = $3; - if ($7 != nullptr) { - $$->insertion.values.swap(*$7); - delete $7; + if ($5 != nullptr) { + $$->insertion.values_list.swap(*$5); + delete $5; + } + free($3); + } + | INSERT INTO ID LBRACE attr_list RBRACE VALUES values_list + { + $$ = new ParsedSqlNode(SCF_INSERT); + $$->insertion.relation_name = $3; + $$->insertion.attr_names = std::move(*$5); + if ($8 != nullptr) { + $$->insertion.values_list.swap(*$8); + delete $8; } - $$->insertion.values.emplace_back(*$6); - std::reverse($$->insertion.values.begin(), $$->insertion.values.end()); - delete $6; free($3); } ; +values_list: + LBRACE value_list RBRACE + { + $$ = new std::vector>; + $$->emplace_back(*$2); + delete $2; + } + | values_list COMMA LBRACE value_list RBRACE + { + $$->emplace_back(*$4); + delete $4; + } + +digits: + NUMBER + { + $$ = float($1); + } + | '-' NUMBER + { + $$ = float(-$2); + } + | FLOAT + { + $$ = $1; + } + | '-' FLOAT + { + $$ = $2; + } + ; + +digits_list: + /* empty */ + { + $$ = new std::vector(); + } + | digits + { + $$ = new std::vector(); + $$->push_back($1); + } + | digits_list COMMA digits + { + $$->push_back($3); + } + ; + value_list: /* empty */ { - $$ = nullptr; + $$ = new std::vector; } - | COMMA value value_list { - if ($3 != nullptr) { - $$ = $3; - } else { - $$ = new std::vector; - } - $$->emplace_back(*$2); - delete $2; + | value + { + $$ = new std::vector; + $$->emplace_back(*$1); + delete $1; + } + | value_list COMMA value + { + $$->emplace_back(*$3); + delete $3; } ; + value: + nonnegative_value { + $$ = $1; + } + | '-' NUMBER { + $$ = new Value(-$2); + @$ = @1; + } + | '-' FLOAT { + $$ = new Value(-$2); + @$ = @1; + } + ; + +nonnegative_value: NUMBER { - $$ = new Value((int)$1); + $$ = new Value($1); @$ = @1; } - |FLOAT { - $$ = new Value((float)$1); + | FLOAT { + $$ = new Value($1); @$ = @1; } - |SSS { + | SSS { char *tmp = common::substr($1,1,strlen($1)-2); $$ = new Value(tmp); free(tmp); free($1); } + | TRUE { + $$ = new Value(true); + } + | FALSE { + $$ = new Value(false); + } + | NULL_T { + $$ = new Value(NullValue()); + } + | LSBRACE digits_list RSBRACE { + $$ = new Value(*$2); + } ; + storage_format: /* empty */ { @@ -427,29 +764,50 @@ delete_stmt: /* delete 语句的语法解析树*/ $$ = new ParsedSqlNode(SCF_DELETE); $$->deletion.relation_name = $3; if ($4 != nullptr) { - $$->deletion.conditions.swap(*$4); - delete $4; + $$->deletion.condition = std::unique_ptr($4); } free($3); } ; + update_stmt: /* update 语句的语法解析树*/ - UPDATE ID SET ID EQ value where + UPDATE ID SET set_clauses where { $$ = new ParsedSqlNode(SCF_UPDATE); $$->update.relation_name = $2; - $$->update.attribute_name = $4; - $$->update.value = *$6; - if ($7 != nullptr) { - $$->update.conditions.swap(*$7); - delete $7; + $$->update.set_clauses.swap(*$4); + if ($5 != nullptr) { + $$->update.conditions = std::unique_ptr($5); } free($2); - free($4); + delete $4; + } + ; + +set_clauses: + set_clause + { + $$ = new std::vector; + $$->emplace_back(std::move(*$1)); + } + | set_clauses COMMA set_clause + { + $$->emplace_back(std::move(*$3)); + } + ; + +set_clause: + ID EQ expression + { + $$ = new SetClauseSqlNode; + $$->field_name = $1; + $$->value = std::unique_ptr($3); + free($1); } ; -select_stmt: /* select 语句的语法解析树*/ - SELECT expression_list FROM rel_list where group_by + +select_stmt: + SELECT expression_list FROM rel_list where group_by opt_having opt_order_by opt_limit { $$ = new ParsedSqlNode(SCF_SELECT); if ($2 != nullptr) { @@ -462,17 +820,63 @@ select_stmt: /* select 语句的语法解析树*/ delete $4; } + $$->selection.conditions = nullptr; + if ($5 != nullptr) { - $$->selection.conditions.swap(*$5); - delete $5; + $$->selection.conditions = std::unique_ptr($5); } if ($6 != nullptr) { $$->selection.group_by.swap(*$6); delete $6; } + + if ($7 != nullptr) { + $$->selection.having_conditions = std::unique_ptr($7); + } + + if ($8 != nullptr) { + $$->selection.order_by.swap(*$8); + delete $8; + } + + if ($9 != nullptr) { + $$->selection.limit = std::make_unique(*$9); + delete $9; + } + } + | SELECT expression_list FROM relation INNER JOIN join_clauses where group_by + { + $$ = new ParsedSqlNode(SCF_SELECT); + if ($2 != nullptr) { + $$->selection.expressions.swap(*$2); + delete $2; + } + + if ($4 != nullptr) { + $$->selection.relations.emplace_back($4); + free($4); + } + + if ($7 != nullptr) { + for (auto it = $7->relations.rbegin(); it != $7->relations.rend(); ++it) { + $$->selection.relations.emplace_back(std::move(*it)); + } + $$->selection.conditions = std::move($7->conditions); + } + + if ($8 != nullptr) { + auto ptr = $$->selection.conditions.release(); + $$->selection.conditions = std::make_unique(ConjunctionExpr::Type::AND, ptr, $8); + } + + if ($9 != nullptr) { + $$->selection.group_by.swap(*$9); + delete $9; + } } ; + calc_stmt: CALC expression_list { @@ -480,24 +884,42 @@ calc_stmt: $$->calc.expressions.swap(*$2); delete $2; } + | SELECT expression_list + { + $$ = new ParsedSqlNode(SCF_CALC); + $$->calc.expressions.swap(*$2); + delete $2; + } ; expression_list: - expression + /* empty */ { + $$ = new std::vector>; + } + | expression alias { $$ = new std::vector>; + if (nullptr != $2) { + $1->set_alias($2); + } $$->emplace_back($1); + free($2); } - | expression COMMA expression_list + | expression alias COMMA expression_list { - if ($3 != nullptr) { - $$ = $3; + if ($4 != nullptr) { + $$ = $4; } else { $$ = new std::vector>; } - $$->emplace($$->begin(), $1); + if (nullptr != $2) { + $1->set_alias($2); + } + $$->emplace($$->begin(),std::move($1)); + free($2); } ; + expression: expression '+' expression { $$ = create_arithmetic_expression(ArithmeticExpr::Type::ADD, $1, $3, sql_string, &@$); @@ -511,14 +933,18 @@ expression: | expression '/' expression { $$ = create_arithmetic_expression(ArithmeticExpr::Type::DIV, $1, $3, sql_string, &@$); } - | LBRACE expression RBRACE { - $$ = $2; + | LBRACE expression_list RBRACE { + if ($2->size() == 1) { + $$ = $2->front().get(); + } else { + $$ = new ListExpr(std::move(*$2)); + } $$->set_name(token_name(sql_string, &@$)); } | '-' expression %prec UMINUS { $$ = create_arithmetic_expression(ArithmeticExpr::Type::NEGATIVE, $2, nullptr, sql_string, &@$); } - | value { + | nonnegative_value { $$ = new ValueExpr(*$1); $$->set_name(token_name(sql_string, &@$)); delete $1; @@ -532,9 +958,44 @@ expression: | '*' { $$ = new StarExpr(); } + | ID DOT '*' { + $$ = new StarExpr($1); + } + | func_expr { + $$ = $1; // AggrFuncExpr + } + | sub_query_expr { + $$ = $1; // SubQueryExpr + } // your code here ; +alias: + /* empty */ { + $$ = nullptr; + } + | ID { + $$ = $1; + } + | AS ID { + $$ = $2; + } + +func_expr: + ID LBRACE expression_list RBRACE + { + $$ = new UnboundFunctionExpr($1, std::move(*$3)); + $$->set_name(token_name(sql_string, &@$)); + } + ; + +sub_query_expr: + LBRACE select_stmt RBRACE + { + $$ = new SubQueryExpr($2->selection); + } + ; + rel_attr: ID { $$ = new RelAttrSqlNode; @@ -555,97 +1016,81 @@ relation: $$ = $1; } ; + rel_list: - relation { - $$ = new std::vector(); - $$->push_back($1); + relation alias { + $$ = new std::vector(); + if(nullptr!=$2){ + $$->emplace_back($1,$2); + free($2); + }else{ + $$->emplace_back($1); + } free($1); } - | relation COMMA rel_list { - if ($3 != nullptr) { - $$ = $3; + | relation alias COMMA rel_list { + if ($4 != nullptr) { + $$ = $4; } else { - $$ = new std::vector; + $$ = new std::vector; + } + if(nullptr!=$2){ + $$->insert($$->begin(), RelationNode($1,$2)); + free($2); + }else{ + $$->insert($$->begin(), RelationNode($1)); } - - $$->insert($$->begin(), $1); free($1); } ; -where: - /* empty */ +join_clauses: + relation ON condition { - $$ = nullptr; + $$ = new JoinSqlNode; + $$->relations.emplace_back($1); + $$->conditions = std::unique_ptr($3); + free($1); } - | WHERE condition_list { - $$ = $2; + | relation ON condition INNER JOIN join_clauses + { + $$ = $6; + $$->relations.emplace_back($1); + auto ptr = $$->conditions.release(); + $$->conditions = std::make_unique(ConjunctionExpr::Type::AND, ptr, $3); + free($1); } ; -condition_list: + +where: /* empty */ { $$ = nullptr; } - | condition { - $$ = new std::vector; - $$->emplace_back(*$1); - delete $1; - } - | condition AND condition_list { - $$ = $3; - $$->emplace_back(*$1); - delete $1; + | WHERE condition { + $$ = $2; } ; + condition: - rel_attr comp_op value + expression comp_op expression { - $$ = new ConditionSqlNode; - $$->left_is_attr = 1; - $$->left_attr = *$1; - $$->right_is_attr = 0; - $$->right_value = *$3; - $$->comp = $2; - - delete $1; - delete $3; + $$ = new ComparisonExpr($2, $1, $3); } - | value comp_op value + | comp_op expression { - $$ = new ConditionSqlNode; - $$->left_is_attr = 0; - $$->left_value = *$1; - $$->right_is_attr = 0; - $$->right_value = *$3; - $$->comp = $2; - - delete $1; - delete $3; + Value val; + val.set_null(true); + ValueExpr *temp_expr = new ValueExpr(val); + $$ = new ComparisonExpr($1,temp_expr, $2); } - | rel_attr comp_op rel_attr + | condition AND condition { - $$ = new ConditionSqlNode; - $$->left_is_attr = 1; - $$->left_attr = *$1; - $$->right_is_attr = 1; - $$->right_attr = *$3; - $$->comp = $2; - - delete $1; - delete $3; + $$ = new ConjunctionExpr(ConjunctionExpr::Type::AND, $1, $3); } - | value comp_op rel_attr + | condition OR condition { - $$ = new ConditionSqlNode; - $$->left_is_attr = 0; - $$->left_value = *$1; - $$->right_is_attr = 1; - $$->right_attr = *$3; - $$->comp = $2; - - delete $1; - delete $3; + $$ = new ConjunctionExpr(ConjunctionExpr::Type::OR, $1, $3); } ; @@ -656,25 +1101,93 @@ comp_op: | LE { $$ = LESS_EQUAL; } | GE { $$ = GREAT_EQUAL; } | NE { $$ = NOT_EQUAL; } + | IS { $$ = IS_OP; } + | IS NOT { $$ = IS_NOT_OP; } + | LIKE { $$ = LIKE_OP;} + | NOT LIKE {$$ = NOT_LIKE_OP;} + | IN { $$ = IN_OP; } + | NOT IN { $$ = NOT_IN_OP; } + | EXISTS { $$ = EXISTS_OP; } + | NOT EXISTS { $$ = NOT_EXISTS_OP; } ; -// your code here +opt_order_by: + /* empty */ + { + $$ = nullptr; + } + | ORDER BY sort_list + { + $$ = $3; + std::reverse($$->begin(),$$->end()); + } + ; + +sort_list: + sort_unit + { + $$ = new std::vector; + $$->emplace_back(std::move(*$1)); + } + | sort_unit COMMA sort_list + { + $3->emplace_back(std::move(*$1)); + $$ = $3; + } + ; + +sort_unit: + expression + { + $$ = new OrderBySqlNode(); + $$->expr = std::unique_ptr($1); + $$->is_asc = true; + } + | expression DESC + { + $$ = new OrderBySqlNode(); + $$->expr = std::unique_ptr($1); + $$->is_asc = false; + } + | expression ASC + { + $$ = new OrderBySqlNode(); // 默认是升序 + $$->expr = std::unique_ptr($1); + $$->is_asc = true; + } + ; + group_by: /* empty */ { $$ = nullptr; } + | GROUP BY expression_list + { + $$ = $3; + } ; -load_data_stmt: - LOAD DATA INFILE SSS INTO TABLE ID + +opt_having: + /* empty */ + { + $$ = nullptr; + } + | HAVING condition + { + $$ = $2; + } + ; + +opt_limit: + /* empty */ + { + $$ = nullptr; + } + | LIMIT NUMBER { - char *tmp_file_name = common::substr($4, 1, strlen($4) - 2); - - $$ = new ParsedSqlNode(SCF_LOAD_DATA); - $$->load_data.relation_name = $7; - $$->load_data.file_name = tmp_file_name; - free($7); - free(tmp_file_name); + $$ = new LimitSqlNode(); + $$->number = $2; } ; @@ -697,8 +1210,8 @@ set_variable_stmt: } ; -opt_semicolon: /*empty*/ - | SEMICOLON +opt_semicolon: + SEMICOLON ; %% //_____________________________________________________________________ diff --git a/src/observer/sql/query_cache/query_cache_stage.cpp b/src/observer/sql/query_cache/query_cache_stage.cpp index eeb53b6b..a8a5f9cc 100644 --- a/src/observer/sql/query_cache/query_cache_stage.cpp +++ b/src/observer/sql/query_cache/query_cache_stage.cpp @@ -24,7 +24,4 @@ See the Mulan PSL v2 for more details. */ using namespace common; -RC QueryCacheStage::handle_request(SQLStageEvent *sql_event) -{ - return RC::SUCCESS; -} +RC QueryCacheStage::handle_request(SQLStageEvent *sql_event) { return RC::SUCCESS; } diff --git a/src/observer/sql/stmt/calc_stmt.h b/src/observer/sql/stmt/calc_stmt.h index d7a1b388..845b2f0e 100644 --- a/src/observer/sql/stmt/calc_stmt.h +++ b/src/observer/sql/stmt/calc_stmt.h @@ -19,6 +19,7 @@ See the Mulan PSL v2 for more details. */ #include "common/rc.h" #include "sql/expr/expression.h" #include "sql/stmt/stmt.h" +#include "sql/parser/expression_binder.h" class Db; class Table; @@ -38,8 +39,20 @@ class CalcStmt : public Stmt public: static RC create(CalcSqlNode &calc_sql, Stmt *&stmt) { - CalcStmt *calc_stmt = new CalcStmt(); - calc_stmt->expressions_ = std::move(calc_sql.expressions); + CalcStmt *calc_stmt = new CalcStmt(); + BinderContext binder_context; + vector> bound_expressions; + ExpressionBinder expression_binder(binder_context); + + for (unique_ptr &expression : calc_sql.expressions) { + RC rc = expression_binder.bind_expression(expression, bound_expressions); + if (OB_FAIL(rc)) { + LOG_INFO("bind expression failed. rc=%s", strrc(rc)); + return rc; + } + } + + calc_stmt->expressions_ = std::move(bound_expressions); stmt = calc_stmt; return RC::SUCCESS; } diff --git a/src/observer/sql/stmt/create_index_stmt.cpp b/src/observer/sql/stmt/create_index_stmt.cpp index 93fae663..fc8de69f 100644 --- a/src/observer/sql/stmt/create_index_stmt.cpp +++ b/src/observer/sql/stmt/create_index_stmt.cpp @@ -26,25 +26,32 @@ RC CreateIndexStmt::create(Db *db, const CreateIndexSqlNode &create_index, Stmt stmt = nullptr; const char *table_name = create_index.relation_name.c_str(); - if (is_blank(table_name) || is_blank(create_index.index_name.c_str()) || - is_blank(create_index.attribute_name.c_str())) { - LOG_WARN("invalid argument. db=%p, table_name=%p, index name=%s, attribute name=%s", - db, table_name, create_index.index_name.c_str(), create_index.attribute_name.c_str()); + if (is_blank(table_name) || is_blank(create_index.index_name.c_str())) { + LOG_WARN("invalid argument. db=%p, table_name=%p, index name=%s", + db, table_name, create_index.index_name.c_str()); return RC::INVALID_ARGUMENT; } // check whether the table exists - Table *table = db->find_table(table_name); - if (nullptr == table) { + BaseTable *base_table = db->find_table(table_name); + if (nullptr == base_table) { LOG_WARN("no such table. db=%s, table_name=%s", db->name(), table_name); return RC::SCHEMA_TABLE_NOT_EXIST; } - const FieldMeta *field_meta = table->table_meta().field(create_index.attribute_name.c_str()); - if (nullptr == field_meta) { - LOG_WARN("no such field in table. db=%s, table=%s, field name=%s", - db->name(), table_name, create_index.attribute_name.c_str()); - return RC::SCHEMA_FIELD_NOT_EXIST; + if (base_table->type() != TableType::Table) { + LOG_WARN("table %s is not BASE TABLE. db=%s", + table_name, db->name()); + return RC::CREATE_INDEX_ON_NON_TABLE_TYPE; + } + + Table *table = static_cast
(base_table); + vector field_metas; + RC rc = table->table_meta().get_field_metas(create_index.attribute_name, field_metas); + if (OB_FAIL(rc)) { + LOG_WARN("no such field in table. db=%s, table=%s", + db->name(), table_name); + return rc; } Index *index = table->find_index(create_index.index_name.c_str()); @@ -53,6 +60,38 @@ RC CreateIndexStmt::create(Db *db, const CreateIndexSqlNode &create_index, Stmt return RC::SCHEMA_INDEX_NAME_REPEAT; } - stmt = new CreateIndexStmt(table, field_meta, create_index.index_name); + auto config = create_index.vector_index_config; + + auto index_type = config.index_type; + if (index_type == IndexType::BPlusTreeIndex) { + stmt = new CreateIndexStmt(table, index_type, field_metas, create_index.index_name, create_index.unique); + } else if (index_type == IndexType::VectorIVFFlatIndex) { + // 解析距离度量方法，后续单独提取出去 + auto distance_type_str = common::str_to_lower(config.distance_fn); + NormalFunctionType distance_type; + if (distance_type_str == "l2_distance") { + distance_type = NormalFunctionType::L2_DISTANCE; + } else if (distance_type_str == "inner_product") { + distance_type = NormalFunctionType::INNER_PRODUCT; + } else if (distance_type_str == "cosine_distance") { + distance_type = NormalFunctionType::COSINE_DISTANCE; + } else { + return RC::UNSUPPORTED; + } + + // 没有参数，使用默认参数 + if (config.lists.attr_type() == AttrType::UNDEFINED && config.probes.attr_type() == AttrType::UNDEFINED) { + stmt = new CreateIndexStmt(table, index_type, field_metas, create_index.index_name, distance_type); + } else if (config.lists.attr_type() == AttrType::INTS && config.probes.attr_type() == AttrType::INTS) { + std::vector options(2); + options[0] = config.lists.get_int(); + options[1] = config.probes.get_int(); + stmt = new CreateIndexStmt( + table, index_type, field_metas, create_index.index_name, distance_type, std::move(options)); + } else { + return RC::INVALID_ARGUMENT; + } + } + return RC::SUCCESS; } diff --git a/src/observer/sql/stmt/create_index_stmt.h b/src/observer/sql/stmt/create_index_stmt.h index 6cd9fbbe..37112660 100644 --- a/src/observer/sql/stmt/create_index_stmt.h +++ b/src/observer/sql/stmt/create_index_stmt.h @@ -17,6 +17,8 @@ See the Mulan PSL v2 for more details. */ #include #include "sql/stmt/stmt.h" +#include "storage/table/table_meta.h" +#include "sql/expr/expression.h" struct CreateIndexSqlNode; class Table; @@ -29,23 +31,43 @@ class FieldMeta; class CreateIndexStmt : public Stmt { public: - CreateIndexStmt(Table *table, const FieldMeta *field_meta, const std::string &index_name) - : table_(table), field_meta_(field_meta), index_name_(index_name) + CreateIndexStmt(Table *table, IndexType index_type, const vector &field_meta, + const std::string &index_name, bool unique) + : table_(table), index_type_(index_type), field_meta_(field_meta), index_name_(index_name), unique_(unique) {} - virtual ~CreateIndexStmt() = default; + // 向量索引 + CreateIndexStmt(Table *table, IndexType index_type, const vector &field_meta, + const std::string &index_name, NormalFunctionType distance_type, std::vector options = {}) + : table_(table), + index_type_(index_type), + field_meta_(field_meta), + index_name_(index_name), + distance_type_(distance_type), + options_(std::move(options)) + {} + + ~CreateIndexStmt() override = default; StmtType type() const override { return StmtType::CREATE_INDEX; } - Table *table() const { return table_; } - const FieldMeta *field_meta() const { return field_meta_; } - const std::string &index_name() const { return index_name_; } + Table *table() const { return table_; } + const vector &field_meta() const { return field_meta_; } + const std::string &index_name() const { return index_name_; } + bool unique() const { return unique_; } + IndexType index_type() const { return index_type_; } + NormalFunctionType distance_type() const { return distance_type_; } + const std::vector &options() const { return options_; } public: static RC create(Db *db, const CreateIndexSqlNode &create_index, Stmt *&stmt); private: - Table *table_ = nullptr; - const FieldMeta *field_meta_ = nullptr; - std::string index_name_; + Table *table_ = nullptr; + IndexType index_type_; + vector field_meta_; + std::string index_name_; + bool unique_ = false; + NormalFunctionType distance_type_; + std::vector options_; }; diff --git a/src/observer/sql/stmt/create_table_stmt.cpp b/src/observer/sql/stmt/create_table_stmt.cpp index 30603af6..161f2cba 100644 --- a/src/observer/sql/stmt/create_table_stmt.cpp +++ b/src/observer/sql/stmt/create_table_stmt.cpp @@ -17,7 +17,7 @@ See the Mulan PSL v2 for more details. */ #include "sql/stmt/create_table_stmt.h" #include "event/sql_debug.h" -RC CreateTableStmt::create(Db *db, const CreateTableSqlNode &create_table, Stmt *&stmt) +RC CreateTableStmt::create(Db *db, CreateTableSqlNode &create_table, Stmt *&stmt) { StorageFormat storage_format = StorageFormat::UNKNOWN_FORMAT; if (create_table.storage_format.length() == 0) { @@ -28,12 +28,26 @@ RC CreateTableStmt::create(Db *db, const CreateTableSqlNode &create_table, Stmt if (storage_format == StorageFormat::UNKNOWN_FORMAT) { return RC::INVALID_ARGUMENT; } - stmt = new CreateTableStmt(create_table.relation_name, create_table.attr_infos, storage_format); + SelectStmt *select_stmt = nullptr; + if (create_table.create_table_select) { + Stmt *create_table_select_stmt; + RC rc = SelectStmt::create(db, *create_table.create_table_select, create_table_select_stmt); + if (OB_FAIL(rc)) { + return rc; + } + select_stmt = dynamic_cast(create_table_select_stmt); + if (select_stmt == nullptr) { + return RC::INTERNAL; + } + } + + stmt = new CreateTableStmt(create_table.relation_name, create_table.attr_infos, storage_format, select_stmt); sql_debug("create table statement: table name %s", create_table.relation_name.c_str()); return RC::SUCCESS; } -StorageFormat CreateTableStmt::get_storage_format(const char *format_str) { +StorageFormat CreateTableStmt::get_storage_format(const char *format_str) +{ StorageFormat format = StorageFormat::UNKNOWN_FORMAT; if (0 == strcasecmp(format_str, "ROW")) { format = StorageFormat::ROW_FORMAT; diff --git a/src/observer/sql/stmt/create_table_stmt.h b/src/observer/sql/stmt/create_table_stmt.h index a7c0707e..f74dde7d 100644 --- a/src/observer/sql/stmt/create_table_stmt.h +++ b/src/observer/sql/stmt/create_table_stmt.h @@ -18,6 +18,7 @@ See the Mulan PSL v2 for more details. */ #include #include "sql/stmt/stmt.h" +#include "select_stmt.h" class Db; @@ -29,9 +30,12 @@ class Db; class CreateTableStmt : public Stmt { public: - CreateTableStmt( - const std::string &table_name, const std::vector &attr_infos, StorageFormat storage_format) - : table_name_(table_name), attr_infos_(attr_infos), storage_format_(storage_format) + CreateTableStmt(const std::string &table_name, const std::vector &attr_infos, + StorageFormat storage_format, SelectStmt *create_table_select_stmt) + : table_name_(table_name), + attr_infos_(attr_infos), + storage_format_(storage_format), + create_table_select_stmt_(create_table_select_stmt) {} virtual ~CreateTableStmt() = default; @@ -40,12 +44,14 @@ class CreateTableStmt : public Stmt const std::string &table_name() const { return table_name_; } const std::vector &attr_infos() const { return attr_infos_; } const StorageFormat storage_format() const { return storage_format_; } + SelectStmt *create_table_select_stmt() { return create_table_select_stmt_; } - static RC create(Db *db, const CreateTableSqlNode &create_table, Stmt *&stmt); + static RC create(Db *db, CreateTableSqlNode &create_table, Stmt *&stmt); static StorageFormat get_storage_format(const char *format_str); private: std::string table_name_; std::vector attr_infos_; StorageFormat storage_format_; + SelectStmt *create_table_select_stmt_; }; \ No newline at end of file diff --git a/src/observer/sql/stmt/create_view_stmt.cpp b/src/observer/sql/stmt/create_view_stmt.cpp new file mode 100644 index 00000000..a3b7b46a --- /dev/null +++ b/src/observer/sql/stmt/create_view_stmt.cpp @@ -0,0 +1,43 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "create_view_stmt.h" +#include "sql/stmt/create_table_stmt.h" +#include "event/sql_debug.h" + +RC CreateViewStmt::create(Db *db, CreateViewSqlNode &create_view, Stmt *&stmt) +{ + SelectStmt *select_stmt = nullptr; + if (create_view.create_view_select) { + Stmt *create_view_select_stmt; + RC rc = SelectStmt::create(db, *create_view.create_view_select, create_view_select_stmt); + if (OB_FAIL(rc)) { + return rc; + } + select_stmt = dynamic_cast(create_view_select_stmt); + if (select_stmt == nullptr) { + return RC::INTERNAL; + } + } else { + return RC::INTERNAL; + } + + if (!create_view.attribute_names.empty() && + create_view.attribute_names.size() != select_stmt->query_expressions_size()) { + LOG_ERROR("In definition of view, derived table or common table expression, SELECT list and column names list have different column counts"); + return RC::INVALID_ARGUMENT; + } + + stmt = new CreateViewStmt(create_view.relation_name, create_view.attribute_names, select_stmt); + sql_debug("create view statement: view name %s", create_view.relation_name.c_str()); + return RC::SUCCESS; +} diff --git a/src/observer/sql/stmt/create_view_stmt.h b/src/observer/sql/stmt/create_view_stmt.h new file mode 100644 index 00000000..ae2cee9c --- /dev/null +++ b/src/observer/sql/stmt/create_view_stmt.h @@ -0,0 +1,49 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include +#include +#include + +#include "sql/stmt/stmt.h" +#include "select_stmt.h" + +class Db; + +/** + * @brief 表示创建表的语句 + * @ingroup Statement + * @details 虽然解析成了stmt，但是与原始的SQL解析后的数据也差不多 + */ +class CreateViewStmt : public Stmt +{ +public: + CreateViewStmt(std::string &table_name, std::vector &attr_names, SelectStmt *select_stmt) + : table_name_(std::move(table_name)), attr_names_(std::move(attr_names)), select_stmt_(select_stmt) + {} + ~CreateViewStmt() override = default; + + StmtType type() const override { return StmtType::CREATE_VIEW; } + + std::string &table_name() { return table_name_; } + std::vector &attr_names() { return attr_names_; } + SelectStmt *select_stmt() { return select_stmt_; } + + static RC create(Db *db, CreateViewSqlNode &create_view, Stmt *&stmt); + +private: + std::string table_name_; + std::vector attr_names_; // 是否指定了属性名 + SelectStmt *select_stmt_; +}; diff --git a/src/observer/sql/stmt/delete_stmt.cpp b/src/observer/sql/stmt/delete_stmt.cpp index 36f8c1f8..76a39d80 100644 --- a/src/observer/sql/stmt/delete_stmt.cpp +++ b/src/observer/sql/stmt/delete_stmt.cpp @@ -17,8 +17,9 @@ See the Mulan PSL v2 for more details. */ #include "sql/stmt/filter_stmt.h" #include "storage/db/db.h" #include "storage/table/table.h" +#include "storage/table/view.h" -DeleteStmt::DeleteStmt(Table *table, FilterStmt *filter_stmt) : table_(table), filter_stmt_(filter_stmt) {} +DeleteStmt::DeleteStmt(BaseTable *table, FilterStmt *filter_stmt) : table_(table), filter_stmt_(filter_stmt) {} DeleteStmt::~DeleteStmt() { @@ -28,7 +29,7 @@ DeleteStmt::~DeleteStmt() } } -RC DeleteStmt::create(Db *db, const DeleteSqlNode &delete_sql, Stmt *&stmt) +RC DeleteStmt::create(Db *db, DeleteSqlNode &delete_sql, Stmt *&stmt) { const char *table_name = delete_sql.relation_name.c_str(); if (nullptr == db || nullptr == table_name) { @@ -37,18 +38,29 @@ RC DeleteStmt::create(Db *db, const DeleteSqlNode &delete_sql, Stmt *&stmt) } // check whether the table exists - Table *table = db->find_table(table_name); + BaseTable *table = db->find_table(table_name); if (nullptr == table) { LOG_WARN("no such table. db=%s, table_name=%s", db->name(), table_name); return RC::SCHEMA_TABLE_NOT_EXIST; } - std::unordered_map table_map; - table_map.insert(std::pair(std::string(table_name), table)); + if (table->type() == TableType::View) { + if (!table->is_mutable()) { + LOG_ERROR("The target table %s of the DELETE is not updatable", table->name()); + return RC::READ_ONLY_VIEW_DELETE_ERROR; + } + auto view = dynamic_cast(table); + if (view->has_join()) { + LOG_ERROR("Can not delete from join view '%s.%s' without fields list", db->name(), table->name()); + return RC::JOIN_VIEW_DELETE_ERROR; + } + } + + std::unordered_map table_map; + table_map.insert(std::pair(std::string(table_name), table)); FilterStmt *filter_stmt = nullptr; - RC rc = FilterStmt::create( - db, table, &table_map, delete_sql.conditions.data(), static_cast(delete_sql.conditions.size()), filter_stmt); + RC rc = FilterStmt::create(db, table, {}, &table_map, delete_sql.condition, filter_stmt); if (rc != RC::SUCCESS) { LOG_WARN("failed to create filter statement. rc=%d:%s", rc, strrc(rc)); return rc; diff --git a/src/observer/sql/stmt/delete_stmt.h b/src/observer/sql/stmt/delete_stmt.h index 33cda12a..0d3a6300 100644 --- a/src/observer/sql/stmt/delete_stmt.h +++ b/src/observer/sql/stmt/delete_stmt.h @@ -17,7 +17,7 @@ See the Mulan PSL v2 for more details. */ #include "sql/parser/parse_defs.h" #include "sql/stmt/stmt.h" -class Table; +class BaseTable; class FilterStmt; /** @@ -27,18 +27,18 @@ class FilterStmt; class DeleteStmt : public Stmt { public: - DeleteStmt(Table *table, FilterStmt *filter_stmt); + DeleteStmt(BaseTable *table, FilterStmt *filter_stmt); ~DeleteStmt() override; - Table *table() const { return table_; } + BaseTable *table() const { return table_; } FilterStmt *filter_stmt() const { return filter_stmt_; } StmtType type() const override { return StmtType::DELETE; } public: - static RC create(Db *db, const DeleteSqlNode &delete_sql, Stmt *&stmt); + static RC create(Db *db, DeleteSqlNode &delete_sql, Stmt *&stmt); private: - Table *table_ = nullptr; + BaseTable *table_ = nullptr; FilterStmt *filter_stmt_ = nullptr; }; diff --git a/src/observer/sql/stmt/drop_table_stmt.cpp b/src/observer/sql/stmt/drop_table_stmt.cpp new file mode 100644 index 00000000..1b26dedd --- /dev/null +++ b/src/observer/sql/stmt/drop_table_stmt.cpp @@ -0,0 +1,22 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/15 * + * @Description : DropTableStmt source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "drop_table_stmt.h" + +#include "event/sql_debug.h" + +RC DropTableStmt::create(Db *db, const DropTableSqlNode &drop_table, Stmt *&stmt) +{ + stmt = new DropTableStmt(drop_table.relation_name); + sql_debug("drop table statement: table name %s", drop_table.relation_name.c_str()); + return RC::SUCCESS; +} diff --git a/src/observer/sql/stmt/drop_table_stmt.h b/src/observer/sql/stmt/drop_table_stmt.h new file mode 100644 index 00000000..4acaa90e --- /dev/null +++ b/src/observer/sql/stmt/drop_table_stmt.h @@ -0,0 +1,41 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/15 * + * @Description : DropTableStmt header file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include +#include + +#include "sql/stmt/stmt.h" + +class Db; + +/** + * @brief 表示删除表的语句 + * @ingroup Statement + * @details 虽然解析成了stmt，但是与原始的SQL解析后的数据也差不多 + */ +class DropTableStmt : public Stmt +{ +public: + DropTableStmt(const std::string &table_name) : table_name_(table_name) {} + virtual ~DropTableStmt() = default; + + StmtType type() const override { return StmtType::DROP_TABLE; } + + const std::string &table_name() const { return table_name_; } + + static RC create(Db *db, const DropTableSqlNode &drop_table, Stmt *&stmt); + +private: + std::string table_name_; +}; diff --git a/src/observer/sql/stmt/drop_view_stmt.cpp b/src/observer/sql/stmt/drop_view_stmt.cpp new file mode 100644 index 00000000..607e6038 --- /dev/null +++ b/src/observer/sql/stmt/drop_view_stmt.cpp @@ -0,0 +1,21 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "drop_view_stmt.h" +#include "event/sql_debug.h" + +RC DropViewStmt::create(Db *db, const DropViewSqlNode &drop_view, Stmt *&stmt) +{ + stmt = new DropViewStmt(drop_view.relation_name); + sql_debug("drop view statement: view name %s", drop_view.relation_name.c_str()); + return RC::SUCCESS; +} diff --git a/src/observer/sql/stmt/drop_view_stmt.h b/src/observer/sql/stmt/drop_view_stmt.h new file mode 100644 index 00000000..5bc018fb --- /dev/null +++ b/src/observer/sql/stmt/drop_view_stmt.h @@ -0,0 +1,41 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include +#include + +#include "sql/stmt/stmt.h" + +class Db; + +/** + * @brief 表示删除视图的语句 + * @ingroup Statement + * @details 虽然解析成了stmt，但是与原始的SQL解析后的数据也差不多 + */ +class DropViewStmt : public Stmt +{ +public: + explicit DropViewStmt(std::string table_name) : table_name_(std::move(table_name)) {} + ~DropViewStmt() override = default; + + StmtType type() const override { return StmtType::DROP_VIEW; } + + const std::string &table_name() const { return table_name_; } + + static RC create(Db *db, const DropViewSqlNode &drop_view, Stmt *&stmt); + +private: + std::string table_name_; +}; diff --git a/src/observer/sql/stmt/filter_stmt.cpp b/src/observer/sql/stmt/filter_stmt.cpp index 44b1d1a1..33edee9c 100644 --- a/src/observer/sql/stmt/filter_stmt.cpp +++ b/src/observer/sql/stmt/filter_stmt.cpp @@ -19,114 +19,42 @@ See the Mulan PSL v2 for more details. */ #include "storage/db/db.h" #include "storage/table/table.h" -FilterStmt::~FilterStmt() -{ - for (FilterUnit *unit : filter_units_) { - delete unit; - } - filter_units_.clear(); -} +FilterStmt::~FilterStmt() = default; -RC FilterStmt::create(Db *db, Table *default_table, std::unordered_map *tables, - const ConditionSqlNode *conditions, int condition_num, FilterStmt *&stmt) +RC FilterStmt::create(Db *db, BaseTable *default_table, std::vector tables_alias, + std::unordered_map *tables, std::unique_ptr &condition, FilterStmt *&stmt) { RC rc = RC::SUCCESS; stmt = nullptr; - - FilterStmt *tmp_stmt = new FilterStmt(); - for (int i = 0; i < condition_num; i++) { - FilterUnit *filter_unit = nullptr; - - rc = create_filter_unit(db, default_table, tables, conditions[i], filter_unit); - if (rc != RC::SUCCESS) { - delete tmp_stmt; - LOG_WARN("failed to create filter unit. condition index=%d", i); - return rc; - } - tmp_stmt->filter_units_.push_back(filter_unit); + // 1、 没有条件直接返回 + if (nullptr == condition) { + return rc; } - stmt = tmp_stmt; - return rc; -} + // 2、 检查条件的合法性 + vector> cond_expressions; + BinderContext binder_context; -RC get_table_and_field(Db *db, Table *default_table, std::unordered_map *tables, - const RelAttrSqlNode &attr, Table *&table, const FieldMeta *&field) -{ - if (common::is_blank(attr.relation_name.c_str())) { - table = default_table; - } else if (nullptr != tables) { - auto iter = tables->find(attr.relation_name); - if (iter != tables->end()) { - table = iter->second; + if (tables != nullptr) { + for (const auto &i : *tables) { + binder_context.add_table(i.second); } - } else { - table = db->find_table(attr.relation_name.c_str()); - } - if (nullptr == table) { - LOG_WARN("No such table: attr.relation_name: %s", attr.relation_name.c_str()); - return RC::SCHEMA_TABLE_NOT_EXIST; - } - - field = table->table_meta().field(attr.attribute_name.c_str()); - if (nullptr == field) { - LOG_WARN("no such field in table: table %s, field %s", table->name(), attr.attribute_name.c_str()); - table = nullptr; - return RC::SCHEMA_FIELD_NOT_EXIST; } + binder_context.add_db(db); + binder_context.set_alias(std::move(tables_alias)); + binder_context.set_tables(tables); + binder_context.set_default_table(default_table); + ExpressionBinder expression_binder(binder_context); - return RC::SUCCESS; -} - -RC FilterStmt::create_filter_unit(Db *db, Table *default_table, std::unordered_map *tables, - const ConditionSqlNode &condition, FilterUnit *&filter_unit) -{ - RC rc = RC::SUCCESS; - - CompOp comp = condition.comp; - if (comp < EQUAL_TO || comp >= NO_OP) { - LOG_WARN("invalid compare operator : %d", comp); - return RC::INVALID_ARGUMENT; - } - - filter_unit = new FilterUnit; + rc = expression_binder.bind_expression(condition, cond_expressions); - if (condition.left_is_attr) { - Table *table = nullptr; - const FieldMeta *field = nullptr; - rc = get_table_and_field(db, default_table, tables, condition.left_attr, table, field); - if (rc != RC::SUCCESS) { - LOG_WARN("cannot find attr"); - return rc; - } - FilterObj filter_obj; - filter_obj.init_attr(Field(table, field)); - filter_unit->set_left(filter_obj); - } else { - FilterObj filter_obj; - filter_obj.init_value(condition.left_value); - filter_unit->set_left(filter_obj); - } - - if (condition.right_is_attr) { - Table *table = nullptr; - const FieldMeta *field = nullptr; - rc = get_table_and_field(db, default_table, tables, condition.right_attr, table, field); - if (rc != RC::SUCCESS) { - LOG_WARN("cannot find attr"); - return rc; - } - FilterObj filter_obj; - filter_obj.init_attr(Field(table, field)); - filter_unit->set_right(filter_obj); - } else { - FilterObj filter_obj; - filter_obj.init_value(condition.right_value); - filter_unit->set_right(filter_obj); + if (OB_FAIL(rc)) { + LOG_WARN("CAN NOT PASS CONDITION CHECK!"); + return rc; } + FilterStmt *tmp_stmt = new FilterStmt(); + tmp_stmt->condition_ = std::move(cond_expressions[0]); - filter_unit->set_comp(comp); - - // 检查两个类型是否能够比较 + stmt = tmp_stmt; return rc; } diff --git a/src/observer/sql/stmt/filter_stmt.h b/src/observer/sql/stmt/filter_stmt.h index 27017782..e298e164 100644 --- a/src/observer/sql/stmt/filter_stmt.h +++ b/src/observer/sql/stmt/filter_stmt.h @@ -17,6 +17,7 @@ See the Mulan PSL v2 for more details. */ #include "sql/expr/expression.h" #include "sql/parser/parse_defs.h" #include "sql/stmt/stmt.h" +#include "sql/parser/expression_binder.h" #include #include @@ -24,47 +25,6 @@ class Db; class Table; class FieldMeta; -struct FilterObj -{ - bool is_attr; - Field field; - Value value; - - void init_attr(const Field &field) - { - is_attr = true; - this->field = field; - } - - void init_value(const Value &value) - { - is_attr = false; - this->value = value; - } -}; - -class FilterUnit -{ -public: - FilterUnit() = default; - ~FilterUnit() {} - - void set_comp(CompOp comp) { comp_ = comp; } - - CompOp comp() const { return comp_; } - - void set_left(const FilterObj &obj) { left_ = obj; } - void set_right(const FilterObj &obj) { right_ = obj; } - - const FilterObj &left() const { return left_; } - const FilterObj &right() const { return right_; } - -private: - CompOp comp_ = NO_OP; - FilterObj left_; - FilterObj right_; -}; - /** * @brief Filter/谓词/过滤语句 * @ingroup Statement @@ -76,15 +36,16 @@ class FilterStmt virtual ~FilterStmt(); public: - const std::vector &filter_units() const { return filter_units_; } + static RC create(Db *db, BaseTable *default_table, std::vector tables_alias, + std::unordered_map *tables, std::unique_ptr &condition, FilterStmt *&stmt); -public: - static RC create(Db *db, Table *default_table, std::unordered_map *tables, - const ConditionSqlNode *conditions, int condition_num, FilterStmt *&stmt); + bool condition_empty() const { return nullptr == condition_; } + std::unique_ptr &condition() { return condition_; } - static RC create_filter_unit(Db *db, Table *default_table, std::unordered_map *tables, - const ConditionSqlNode &condition, FilterUnit *&filter_unit); +private: + RC bind_expressions_recursively(ExpressionBinder &expression_binder, unique_ptr &condition, + vector> &cond_expressions); private: - std::vector filter_units_; // 默认当前都是AND关系 + std::unique_ptr condition_; }; diff --git a/src/observer/sql/stmt/insert_stmt.cpp b/src/observer/sql/stmt/insert_stmt.cpp index 96ced424..2e0f8933 100644 --- a/src/observer/sql/stmt/insert_stmt.cpp +++ b/src/observer/sql/stmt/insert_stmt.cpp @@ -12,42 +12,110 @@ See the Mulan PSL v2 for more details. */ // Created by Wangyunlai on 2022/5/22. // -#include "sql/stmt/insert_stmt.h" +#include + #include "common/log/log.h" #include "storage/db/db.h" #include "storage/table/table.h" +#include "storage/table/view.h" +#include "sql/stmt/insert_stmt.h" -InsertStmt::InsertStmt(Table *table, const Value *values, int value_amount) - : table_(table), values_(values), value_amount_(value_amount) +InsertStmt::InsertStmt(BaseTable *table, std::vector> values_list) + : table_(table), values_list_(std::move(values_list)) {} RC InsertStmt::create(Db *db, const InsertSqlNode &inserts, Stmt *&stmt) { const char *table_name = inserts.relation_name.c_str(); - if (nullptr == db || nullptr == table_name || inserts.values.empty()) { + if (nullptr == db || nullptr == table_name || inserts.values_list.empty()) { LOG_WARN("invalid argument. db=%p, table_name=%p, value_num=%d", - db, table_name, static_cast(inserts.values.size())); + db, table_name, static_cast(inserts.values_list.size())); return RC::INVALID_ARGUMENT; } // check whether the table exists - Table *table = db->find_table(table_name); + auto table = db->find_table(table_name); if (nullptr == table) { LOG_WARN("no such table. db=%s, table_name=%s", db->name(), table_name); return RC::SCHEMA_TABLE_NOT_EXIST; } + if (table->type() == TableType::View) { + if (!table->is_mutable()) { + LOG_ERROR("The target table %s of the INSERT is not insertable-into", table->name()); + return RC::READ_ONLY_VIEW_INSERT_ERROR; + } + auto view = dynamic_cast(table); + if (view->has_join()) { + LOG_ERROR("Can not insert into join view '%s.%s' without fields list", db->name(), table->name()); + return RC::JOIN_VIEW_INSERT_ERROR; + } + } + + // check the fields are mutable + // 有表达式的列整个表都不能插入，但是可以删除更新非表达式列 + const TableMeta &table_meta = table->table_meta(); + auto field_metas = table_meta.field_metas(); + for (auto &field_meta : *field_metas) { + if (!field_meta.is_mutable()) { + LOG_ERROR("Column '%s' is not insertable", field_meta.name()); + return RC::EXPRESSION_FIELD_NOT_INSERTABLE; + } + } + + std::vector> values_list = inserts.values_list; + const int field_num = table_meta.field_num() - table_meta.sys_field_num(); + // check the fields number - const Value *values = inserts.values.data(); - const int value_num = static_cast(inserts.values.size()); - const TableMeta &table_meta = table->table_meta(); - const int field_num = table_meta.field_num() - table_meta.sys_field_num(); - if (field_num != value_num) { - LOG_WARN("schema mismatch. value num=%d, field num in schema=%d", value_num, field_num); - return RC::SCHEMA_FIELD_MISSING; + for (auto &value_list : inserts.values_list) { + const int value_num = static_cast(value_list.size()); + if (inserts.attr_names.empty()) { + if (field_num != value_num) { + LOG_WARN("schema mismatch. value num=%d, field num in schema=%d", value_num, field_num); + return RC::SCHEMA_FIELD_MISSING; + } + } else if (field_num < value_num || inserts.attr_names.size() != value_num) { + LOG_WARN("schema mismatch. attr num=%d, value num=%d, field num in schema=%d", inserts.attr_names.size(), value_num, field_num); + return RC::INVALID_ARGUMENT; + } + } + + if (!inserts.attr_names.empty()) { + // 在物理算子执行阶段检查值的可为空性 + // 预处理索引 + std::unordered_set field_ids; + std::vector index(inserts.attr_names.size(), -1); + for (int i = 0; i < inserts.attr_names.size(); ++i) { + auto &attr_name = inserts.attr_names[i]; + auto field_meta = table_meta.field(attr_name.c_str()); + if (field_meta != nullptr) { + // 出现两次同名列 + if (field_ids.count(field_meta->field_id())) { + LOG_ERROR("Column '%s' specified twice", attr_name.c_str()); + return RC::INVALID_ARGUMENT; + } + index[i] = field_meta->field_id(); + field_ids.emplace(index[i]); + } else { + LOG_WARN("Field does not exist. db=%s, table_name=%s, field_name=%s", + db->name(), table_name, attr_name.c_str()); + return RC::SCHEMA_FIELD_NOT_EXIST; + } + } + + for (auto &value_list : values_list) { + std::vector values(field_num); + for (auto &value : values) { + value.set_null(); + } + for (int k = 0; k < value_list.size(); ++k) { + values[index[k]] = value_list[k]; + } + value_list = std::move(values); + } } // everything alright - stmt = new InsertStmt(table, values, value_num); + stmt = new InsertStmt(table, std::move(values_list)); return RC::SUCCESS; } diff --git a/src/observer/sql/stmt/insert_stmt.h b/src/observer/sql/stmt/insert_stmt.h index 759745cb..29b87e65 100644 --- a/src/observer/sql/stmt/insert_stmt.h +++ b/src/observer/sql/stmt/insert_stmt.h @@ -17,7 +17,7 @@ See the Mulan PSL v2 for more details. */ #include "common/rc.h" #include "sql/stmt/stmt.h" -class Table; +class BaseTable; class Db; /** @@ -27,21 +27,17 @@ class Db; class InsertStmt : public Stmt { public: - InsertStmt() = default; - InsertStmt(Table *table, const Value *values, int value_amount); + InsertStmt() = delete; + InsertStmt(BaseTable *table, std::vector> values_list); StmtType type() const override { return StmtType::INSERT; } -public: static RC create(Db *db, const InsertSqlNode &insert_sql, Stmt *&stmt); -public: - Table *table() const { return table_; } - const Value *values() const { return values_; } - int value_amount() const { return value_amount_; } + BaseTable *table() const { return table_; } + const std::vector> &values_list() const { return values_list_; }; private: - Table *table_ = nullptr; - const Value *values_ = nullptr; - int value_amount_ = 0; + BaseTable *table_ = nullptr; + std::vector> values_list_; }; diff --git a/src/observer/sql/stmt/load_data_stmt.cpp b/src/observer/sql/stmt/load_data_stmt.cpp index 530faa42..30a15357 100644 --- a/src/observer/sql/stmt/load_data_stmt.cpp +++ b/src/observer/sql/stmt/load_data_stmt.cpp @@ -32,7 +32,7 @@ RC LoadDataStmt::create(Db *db, const LoadDataSqlNode &load_data, Stmt *&stmt) } // check whether the table exists - Table *table = db->find_table(table_name); + auto table = db->find_table(table_name); if (nullptr == table) { LOG_WARN("no such table. db=%s, table_name=%s", db->name(), table_name); return RC::SCHEMA_TABLE_NOT_EXIST; diff --git a/src/observer/sql/stmt/load_data_stmt.h b/src/observer/sql/stmt/load_data_stmt.h index 9904e82c..742ebfbc 100644 --- a/src/observer/sql/stmt/load_data_stmt.h +++ b/src/observer/sql/stmt/load_data_stmt.h @@ -18,22 +18,22 @@ See the Mulan PSL v2 for more details. */ #include "sql/stmt/stmt.h" -class Table; +class BaseTable; class LoadDataStmt : public Stmt { public: - LoadDataStmt(Table *table, const char *filename) : table_(table), filename_(filename) {} + LoadDataStmt(BaseTable *table, const char *filename) : table_(table), filename_(filename) {} virtual ~LoadDataStmt() = default; StmtType type() const override { return StmtType::LOAD_DATA; } - Table *table() const { return table_; } + BaseTable *table() const { return table_; } const char *filename() const { return filename_.c_str(); } static RC create(Db *db, const LoadDataSqlNode &load_data, Stmt *&stmt); private: - Table *table_ = nullptr; + BaseTable *table_ = nullptr; std::string filename_; }; diff --git a/src/observer/sql/stmt/select_stmt.cpp b/src/observer/sql/stmt/select_stmt.cpp index a51d9e79..eebca17d 100644 --- a/src/observer/sql/stmt/select_stmt.cpp +++ b/src/observer/sql/stmt/select_stmt.cpp @@ -31,7 +31,8 @@ SelectStmt::~SelectStmt() } } -RC SelectStmt::create(Db *db, SelectSqlNode &select_sql, Stmt *&stmt) +RC SelectStmt::create(Db *db, SelectSqlNode &select_sql, Stmt *&stmt, + const std::unordered_map &parent_table_map) { if (nullptr == db) { LOG_WARN("invalid argument. db is null"); @@ -41,30 +42,53 @@ RC SelectStmt::create(Db *db, SelectSqlNode &select_sql, Stmt *&stmt) BinderContext binder_context; // collect tables in `from` statement - vector
tables; - unordered_map table_map; + vector tables; + unordered_map table_map = parent_table_map; + unordered_map temp_map; + std::vector tables_alias(select_sql.relations.size()); + for (size_t i = 0; i < select_sql.relations.size(); i++) { - const char *table_name = select_sql.relations[i].c_str(); + const char *table_name = select_sql.relations[i].relation.c_str(); if (nullptr == table_name) { LOG_WARN("invalid argument. relation name is null. index=%d", i); return RC::INVALID_ARGUMENT; } - Table *table = db->find_table(table_name); + BaseTable *table = db->find_table(table_name); if (nullptr == table) { LOG_WARN("no such table. db=%s, table_name=%s", db->name(), table_name); return RC::SCHEMA_TABLE_NOT_EXIST; } + // 建立别名 + auto &table_alias = select_sql.relations[i].alias; + if (!table_alias.empty()) { + const auto &success = temp_map.emplace(table_alias, table); + if (!success.second) + return RC::INVALID_ALIAS; + } else { + temp_map.emplace(table_name, table); + } + tables_alias[i] = table_alias; + tables.emplace_back(table); binder_context.add_table(table); - tables.push_back(table); - table_map.insert({table_name, table}); } + // alias is all avaliable + table_map.insert(temp_map.begin(), temp_map.end()); + + BaseTable *default_table = nullptr; + if (tables.size() == 1) { + default_table = tables[0]; + } + + binder_context.set_alias(tables_alias); + binder_context.set_tables(&table_map); + binder_context.set_default_table(default_table); // collect query fields in `select` statement vector> bound_expressions; - ExpressionBinder expression_binder(binder_context); - + ExpressionBinder expression_binder(binder_context); + for (unique_ptr &expression : select_sql.expressions) { RC rc = expression_binder.bind_expression(expression, bound_expressions); if (OB_FAIL(rc)) { @@ -82,31 +106,54 @@ RC SelectStmt::create(Db *db, SelectSqlNode &select_sql, Stmt *&stmt) } } - Table *default_table = nullptr; - if (tables.size() == 1) { - default_table = tables[0]; + vector> order_by_expressions; + for (OrderBySqlNode &unit : select_sql.order_by) { + RC rc = expression_binder.bind_expression(unit.expr, order_by_expressions); + if (OB_FAIL(rc)) { + LOG_INFO("bind expression failed. rc=%s", strrc(rc)); + return rc; + } + } + + std::vector order_by_; + order_by_.reserve(order_by_expressions.size()); + for (size_t i = 0; i < order_by_expressions.size(); i++) { + order_by_.push_back({std::move(order_by_expressions[i]), select_sql.order_by[i].is_asc}); + } + + int limit = -1; + if (select_sql.limit) { + limit = select_sql.limit->number; } // create filter statement in `where` statement FilterStmt *filter_stmt = nullptr; - RC rc = FilterStmt::create(db, - default_table, - &table_map, - select_sql.conditions.data(), - static_cast(select_sql.conditions.size()), - filter_stmt); + RC rc = FilterStmt::create(db, default_table, binder_context.alias(), &table_map, select_sql.conditions, filter_stmt); if (rc != RC::SUCCESS) { LOG_WARN("cannot construct filter stmt"); return rc; } + // create filter statement in `having` statement + FilterStmt *having_filter_stmt = nullptr; + rc = FilterStmt::create( + db, default_table, binder_context.alias(), &table_map, select_sql.having_conditions, having_filter_stmt); + if (rc != RC::SUCCESS) { + LOG_WARN("cannot construct having filter stmt"); + return rc; + } + // everything alright SelectStmt *select_stmt = new SelectStmt(); select_stmt->tables_.swap(tables); + select_stmt->tables_alias_ = std::move(tables_alias); select_stmt->query_expressions_.swap(bound_expressions); select_stmt->filter_stmt_ = filter_stmt; select_stmt->group_by_.swap(group_by_expressions); - stmt = select_stmt; + select_stmt->order_by_.swap(order_by_); + select_stmt->limit_ = limit; + select_stmt->having_filter_stmt_ = having_filter_stmt; + stmt = select_stmt; return RC::SUCCESS; } diff --git a/src/observer/sql/stmt/select_stmt.h b/src/observer/sql/stmt/select_stmt.h index cfb04880..bf0b43e4 100644 --- a/src/observer/sql/stmt/select_stmt.h +++ b/src/observer/sql/stmt/select_stmt.h @@ -37,20 +37,30 @@ class SelectStmt : public Stmt ~SelectStmt() override; StmtType type() const override { return StmtType::SELECT; } + size_t query_expressions_size() const { return query_expressions_.size(); } public: - static RC create(Db *db, SelectSqlNode &select_sql, Stmt *&stmt); + static RC create(Db *db, SelectSqlNode &select_sql, Stmt *&stmt, + const std::unordered_map &parent_table_map = {}); public: - const std::vector
&tables() const { return tables_; } - FilterStmt *filter_stmt() const { return filter_stmt_; } + const std::vector &tables() const { return tables_; } + FilterStmt *filter_stmt() const { return filter_stmt_; } + FilterStmt *having_filter_stmt() const { return having_filter_stmt_; } std::vector> &query_expressions() { return query_expressions_; } std::vector> &group_by() { return group_by_; } + std::vector &order_by() { return order_by_; } + std::vector &tables_alias() { return tables_alias_; } + int limit() const { return limit_; } private: std::vector> query_expressions_; - std::vector
tables_; + std::vector tables_; + std::vector tables_alias_; // 存表名 FilterStmt *filter_stmt_ = nullptr; std::vector> group_by_; + std::vector order_by_; + FilterStmt *having_filter_stmt_ = nullptr; + int limit_; }; diff --git a/src/observer/sql/stmt/show_index_stmt.cpp b/src/observer/sql/stmt/show_index_stmt.cpp new file mode 100644 index 00000000..37ce60a4 --- /dev/null +++ b/src/observer/sql/stmt/show_index_stmt.cpp @@ -0,0 +1,25 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/29 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "storage/db/db.h" +#include "sql/stmt/show_index_stmt.h" + +ShowIndexStmt::ShowIndexStmt(const std::string &table_name) : table_name_(table_name) {} + +RC ShowIndexStmt::create(Db *db, const ShowIndexSqlNode &show_index, Stmt *&stmt) +{ + if (db->find_table(show_index.relation_name.c_str()) == nullptr) { + return RC::SCHEMA_TABLE_NOT_EXIST; + } + stmt = new ShowIndexStmt(show_index.relation_name); + return RC::SUCCESS; +} diff --git a/src/observer/sql/stmt/show_index_stmt.h b/src/observer/sql/stmt/show_index_stmt.h new file mode 100644 index 00000000..c5db80d6 --- /dev/null +++ b/src/observer/sql/stmt/show_index_stmt.h @@ -0,0 +1,36 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/9/29 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "sql/stmt/stmt.h" + +/** + * @brief 描述表的语句 + * @ingroup Statement + * @details 虽然解析成了stmt，但是与原始的SQL解析后的数据也差不多 + */ +class ShowIndexStmt : public Stmt +{ +public: + ShowIndexStmt(const std::string &table_name); + ~ShowIndexStmt() override = default; + + StmtType type() const override { return StmtType::SHOW_INDEX; } + + const std::string &table_name() const { return table_name_; } + + static RC create(Db *db, const ShowIndexSqlNode &show_index, Stmt *&stmt); + +private: + const std::string &table_name_; +}; diff --git a/src/observer/sql/stmt/stmt.cpp b/src/observer/sql/stmt/stmt.cpp index 61be9c1b..4b7c61d0 100644 --- a/src/observer/sql/stmt/stmt.cpp +++ b/src/observer/sql/stmt/stmt.cpp @@ -12,8 +12,13 @@ See the Mulan PSL v2 for more details. */ // Created by Wangyunlai on 2022/5/22. // -#include "sql/stmt/stmt.h" #include "common/log/log.h" +#include "drop_table_stmt.h" +#include "sql/stmt/stmt.h" + +#include "create_view_stmt.h" +#include "drop_view_stmt.h" +#include "sql/stmt/update_stmt.h" #include "sql/stmt/calc_stmt.h" #include "sql/stmt/create_index_stmt.h" #include "sql/stmt/create_table_stmt.h" @@ -27,6 +32,7 @@ See the Mulan PSL v2 for more details. */ #include "sql/stmt/select_stmt.h" #include "sql/stmt/set_variable_stmt.h" #include "sql/stmt/show_tables_stmt.h" +#include "sql/stmt/show_index_stmt.h" #include "sql/stmt/trx_begin_stmt.h" #include "sql/stmt/trx_end_stmt.h" @@ -44,6 +50,7 @@ bool stmt_type_ddl(StmtType type) } } } + RC Stmt::create_stmt(Db *db, ParsedSqlNode &sql_node, Stmt *&stmt) { stmt = nullptr; @@ -55,6 +62,9 @@ RC Stmt::create_stmt(Db *db, ParsedSqlNode &sql_node, Stmt *&stmt) case SCF_DELETE: { return DeleteStmt::create(db, sql_node.deletion, stmt); } + case SCF_UPDATE: { + return UpdateStmt::create(db, sql_node.update, stmt); + } case SCF_SELECT: { return SelectStmt::create(db, sql_node.selection, stmt); } @@ -71,6 +81,18 @@ RC Stmt::create_stmt(Db *db, ParsedSqlNode &sql_node, Stmt *&stmt) return CreateTableStmt::create(db, sql_node.create_table, stmt); } + case SCF_DROP_TABLE: { + return DropTableStmt::create(db, sql_node.drop_table, stmt); + } + + case SCF_CREATE_VIEW: { + return CreateViewStmt::create(db, sql_node.create_view, stmt); + } + + case SCF_DROP_VIEW: { + return DropViewStmt::create(db, sql_node.drop_view, stmt); + } + case SCF_DESC_TABLE: { return DescTableStmt::create(db, sql_node.desc_table, stmt); } @@ -83,6 +105,10 @@ RC Stmt::create_stmt(Db *db, ParsedSqlNode &sql_node, Stmt *&stmt) return ShowTablesStmt::create(db, stmt); } + case SCF_SHOW_INDEX: { + return ShowIndexStmt::create(db, sql_node.show_index, stmt); + } + case SCF_BEGIN: { return TrxBeginStmt::create(stmt); } diff --git a/src/observer/sql/stmt/stmt.h b/src/observer/sql/stmt/stmt.h index 2d10ff4c..5df2f0e5 100644 --- a/src/observer/sql/stmt/stmt.h +++ b/src/observer/sql/stmt/stmt.h @@ -39,6 +39,7 @@ class Db; DEFINE_ENUM_ITEM(DROP_TABLE) \ DEFINE_ENUM_ITEM(CREATE_INDEX) \ DEFINE_ENUM_ITEM(DROP_INDEX) \ + DEFINE_ENUM_ITEM(SHOW_INDEX) \ DEFINE_ENUM_ITEM(SYNC) \ DEFINE_ENUM_ITEM(SHOW_TABLES) \ DEFINE_ENUM_ITEM(DESC_TABLE) \ @@ -50,7 +51,9 @@ class Db; DEFINE_ENUM_ITEM(EXIT) \ DEFINE_ENUM_ITEM(EXPLAIN) \ DEFINE_ENUM_ITEM(PREDICATE) \ - DEFINE_ENUM_ITEM(SET_VARIABLE) + DEFINE_ENUM_ITEM(SET_VARIABLE) \ + DEFINE_ENUM_ITEM(CREATE_VIEW) \ + DEFINE_ENUM_ITEM(DROP_VIEW) enum class StmtType { @@ -86,8 +89,5 @@ class Stmt virtual StmtType type() const = 0; -public: static RC create_stmt(Db *db, ParsedSqlNode &sql_node, Stmt *&stmt); - -private: }; diff --git a/src/observer/sql/stmt/update_stmt.cpp b/src/observer/sql/stmt/update_stmt.cpp index 92df6a6d..5e5fc298 100644 --- a/src/observer/sql/stmt/update_stmt.cpp +++ b/src/observer/sql/stmt/update_stmt.cpp @@ -12,15 +12,106 @@ See the Mulan PSL v2 for more details. */ // Created by Wangyunlai on 2022/5/22. // +#include +#include + #include "sql/stmt/update_stmt.h" +#include "sql/stmt/filter_stmt.h" +#include "storage/table/view.h" -UpdateStmt::UpdateStmt(Table *table, Value *values, int value_amount) - : table_(table), values_(values), value_amount_(value_amount) +UpdateStmt::UpdateStmt(BaseTable *table, std::vector field_metas, + std::vector> values, FilterStmt *filter_stmt) + : table_(table), field_metas_(std::move(field_metas)), values_(std::move(values)), filter_stmt_(filter_stmt) {} -RC UpdateStmt::create(Db *db, const UpdateSqlNode &update, Stmt *&stmt) +RC UpdateStmt::create(Db *db, UpdateSqlNode &update_sql, Stmt *&stmt) { // TODO - stmt = nullptr; - return RC::INTERNAL; + const char *table_name = update_sql.relation_name.c_str(); + if (nullptr == db || nullptr == table_name || update_sql.set_clauses.empty()) { + std::ostringstream set_clauses_logger; + set_clauses_logger << "invalid argument. db=" << db << ", table_name=" << table_name; + set_clauses_logger << ", set_clauses=["; + for (const auto &clause : update_sql.set_clauses) { + // 实现表达式打印 + // set_clauses_logger << "{" << clause.field_name << ": " << clause.value.to_string() << "}, "; + set_clauses_logger << "{" << clause.field_name << ": " + << "}, "; + } + set_clauses_logger << "]"; + LOG_WARN("%s", set_clauses_logger.str().c_str()); + return RC::INVALID_ARGUMENT; + } + + // check whether the table exists + auto table = db->find_table(table_name); + if (nullptr == table) { + LOG_WARN("no such table. db=%s, table_name=%s", db->name(), table_name); + return RC::SCHEMA_TABLE_NOT_EXIST; + } + + if (table->type() == TableType::View && !table->is_mutable()) { + LOG_ERROR("The target table %s of the UPDATE is not updatable", table->name()); + return RC::READ_ONLY_VIEW_UPDATE_ERROR; + } + + auto table_meta = table->table_meta(); + std::vector field_metas; + std::vector> values; + + RC rc = RC::SUCCESS; + for (auto &clause : update_sql.set_clauses) { + // check whether the field exists + auto field_meta = table_meta.field(clause.field_name.c_str()); + if (field_meta == nullptr) { + LOG_WARN("Field does not exist. db=%s, table_name=%s, field_name=%s", + db->name(), table_name, clause.field_name.c_str()); + return RC::SCHEMA_FIELD_NOT_EXIST; + } + + if (!field_meta->is_mutable()) { + LOG_ERROR("Column '%s' is not updateable", field_meta->name()); + return RC::EXPRESSION_FIELD_NOT_UPDATEABLE; + } + + // check whether the value is valid + std::unordered_map table_map; + table_map.insert(std::pair(std::string(table_name), table)); + + vector> expressions; + BinderContext binder_context; + + binder_context.add_table(table); + binder_context.add_db(db); + binder_context.set_tables(&table_map); + binder_context.set_default_table(table); + + ExpressionBinder expression_binder(binder_context); + rc = expression_binder.bind_expression(clause.value, expressions); + + if (OB_FAIL(rc)) { + LOG_WARN("Failed to bind expression for field: %s", + clause.field_name.c_str()); + return rc; + } + + // 表达式值类型检查延迟到 UpdatePhysicalOperator 阶段 + + field_metas.emplace_back(*field_meta); + values.emplace_back(std::move(expressions[0])); + } + + std::unordered_map table_map; + table_map.insert(std::pair(std::string(table_name), table)); + + FilterStmt *filter_stmt = nullptr; + rc = FilterStmt::create(db, table, {}, &table_map, update_sql.conditions, filter_stmt); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to create filter statement. rc=%d:%s", rc, strrc(rc)); + return rc; + } + + // everything alright + stmt = new UpdateStmt(table, std::move(field_metas), std::move(values), filter_stmt); + return RC::SUCCESS; } diff --git a/src/observer/sql/stmt/update_stmt.h b/src/observer/sql/stmt/update_stmt.h index 762a8019..c7944afd 100644 --- a/src/observer/sql/stmt/update_stmt.h +++ b/src/observer/sql/stmt/update_stmt.h @@ -16,8 +16,10 @@ See the Mulan PSL v2 for more details. */ #include "common/rc.h" #include "sql/stmt/stmt.h" +#include "storage/field/field_meta.h" -class Table; +class BaseTable; +class FilterStmt; /** * @brief 更新语句 @@ -27,18 +29,21 @@ class UpdateStmt : public Stmt { public: UpdateStmt() = default; - UpdateStmt(Table *table, Value *values, int value_amount); + UpdateStmt(BaseTable *table, std::vector field_metas, std::vector> values, + FilterStmt *filter_stmt); -public: - static RC create(Db *db, const UpdateSqlNode &update_sql, Stmt *&stmt); + StmtType type() const override { return StmtType::UPDATE; } -public: - Table *table() const { return table_; } - Value *values() const { return values_; } - int value_amount() const { return value_amount_; } + BaseTable *table() const { return table_; } + std::vector &field_metas() { return field_metas_; } + std::vector> &values() { return values_; } + FilterStmt *filter_stmt() const { return filter_stmt_; } + + static RC create(Db *db, UpdateSqlNode &update_sql, Stmt *&stmt); private: - Table *table_ = nullptr; - Value *values_ = nullptr; - int value_amount_ = 0; + BaseTable *table_ = nullptr; + std::vector field_metas_; + std::vector> values_; + FilterStmt *filter_stmt_ = nullptr; }; diff --git a/src/observer/storage/buffer/buffer_pool_log.cpp b/src/observer/storage/buffer/buffer_pool_log.cpp index 63b334b8..c8744382 100644 --- a/src/observer/storage/buffer/buffer_pool_log.cpp +++ b/src/observer/storage/buffer/buffer_pool_log.cpp @@ -19,8 +19,7 @@ See the Mulan PSL v2 for more details. */ string BufferPoolLogEntry::to_string() const { - return string("buffer_pool_id=") + std::to_string(buffer_pool_id) + - ", page_num=" + std::to_string(page_num) + + return string("buffer_pool_id=") + std::to_string(buffer_pool_id) + ", page_num=" + std::to_string(page_num) + ", operation_type=" + BufferPoolOperation(operation_type).to_string(); } @@ -40,25 +39,22 @@ RC BufferPoolLogHandler::deallocate_page(PageNum page_num, LSN &lsn) return append_log(BufferPoolOperation::Type::DEALLOCATE, page_num, lsn); } -RC BufferPoolLogHandler::flush_page(Page &page) -{ - return log_handler_.wait_lsn(page.lsn); -} +RC BufferPoolLogHandler::flush_page(Page &page) { return log_handler_.wait_lsn(page.lsn); } RC BufferPoolLogHandler::append_log(BufferPoolOperation::Type type, PageNum page_num, LSN &lsn) { BufferPoolLogEntry log; log.buffer_pool_id = buffer_pool_.id(); - log.page_num = page_num; + log.page_num = page_num; log.operation_type = BufferPoolOperation(type).type_id(); - return log_handler_.append(lsn, LogModule::Id::BUFFER_POOL, span(reinterpret_cast(&log), sizeof(log))); + return log_handler_.append( + lsn, LogModule::Id::BUFFER_POOL, span(reinterpret_cast(&log), sizeof(log))); } //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// // BufferPoolLogReplayer -BufferPoolLogReplayer::BufferPoolLogReplayer(BufferPoolManager &bp_manager) : bp_manager_(bp_manager) -{} +BufferPoolLogReplayer::BufferPoolLogReplayer(BufferPoolManager &bp_manager) : bp_manager_(bp_manager) {} RC BufferPoolLogReplayer::replay(const LogEntry &entry) { @@ -71,10 +67,10 @@ RC BufferPoolLogReplayer::replay(const LogEntry &entry) auto log = reinterpret_cast(entry.data()); LOG_TRACE("replay buffer pool log. entry=%s, log=%s", entry.to_string().c_str(), log->to_string().c_str()); - - int32_t buffer_pool_id = log->buffer_pool_id; - DiskBufferPool *buffer_pool = nullptr; - RC rc = bp_manager_.get_buffer_pool(buffer_pool_id, buffer_pool); + + int32_t buffer_pool_id = log->buffer_pool_id; + DiskBufferPool *buffer_pool = nullptr; + RC rc = bp_manager_.get_buffer_pool(buffer_pool_id, buffer_pool); if (OB_FAIL(rc) || buffer_pool == nullptr) { LOG_ERROR("failed to get buffer pool. rc=%s, buffer pool=%p, log=%s, %s", strrc(rc), buffer_pool, entry.to_string().c_str(), log->to_string().c_str()); @@ -82,12 +78,9 @@ RC BufferPoolLogReplayer::replay(const LogEntry &entry) } BufferPoolOperation operation(log->operation_type); - switch (operation.type()) - { - case BufferPoolOperation::Type::ALLOCATE: - return buffer_pool->redo_allocate_page(entry.lsn(), log->page_num); - case BufferPoolOperation::Type::DEALLOCATE: - return buffer_pool->redo_deallocate_page(entry.lsn(), log->page_num); + switch (operation.type()) { + case BufferPoolOperation::Type::ALLOCATE: return buffer_pool->redo_allocate_page(entry.lsn(), log->page_num); + case BufferPoolOperation::Type::DEALLOCATE: return buffer_pool->redo_deallocate_page(entry.lsn(), log->page_num); default: LOG_ERROR("unknown buffer pool operation. operation=%s", operation.to_string().c_str()); return RC::INTERNAL; diff --git a/src/observer/storage/buffer/disk_buffer_pool.cpp b/src/observer/storage/buffer/disk_buffer_pool.cpp index 92215cc7..2354ba9d 100644 --- a/src/observer/storage/buffer/disk_buffer_pool.cpp +++ b/src/observer/storage/buffer/disk_buffer_pool.cpp @@ -103,7 +103,7 @@ int BPFrameManager::purge_frames(int count, function purger) Frame *BPFrameManager::get(int buffer_pool_id, PageNum page_num) { - FrameId frame_id(buffer_pool_id, page_num); + FrameId frame_id(buffer_pool_id, page_num); lock_guard lock_guard(lock_); return get_internal(frame_id); @@ -125,7 +125,7 @@ Frame *BPFrameManager::alloc(int buffer_pool_id, PageNum page_num) lock_guard lock_guard(lock_); - Frame *frame = get_internal(frame_id); + Frame *frame = get_internal(frame_id); if (frame != nullptr) { return frame; } @@ -170,7 +170,7 @@ list BPFrameManager::find_list(int buffer_pool_id) lock_guard lock_guard(lock_); list frames; - auto fetcher = [&frames, buffer_pool_id](const FrameId &frame_id, Frame *const frame) -> bool { + auto fetcher = [&frames, buffer_pool_id](const FrameId &frame_id, Frame *const frame) -> bool { if (buffer_pool_id == frame_id.buffer_pool_id()) { frame->pin(); frames.push_back(frame); @@ -213,9 +213,12 @@ RC BufferPoolIterator::reset() } //////////////////////////////////////////////////////////////////////////////// -DiskBufferPool::DiskBufferPool( - BufferPoolManager &bp_manager, BPFrameManager &frame_manager, DoubleWriteBuffer &dblwr_manager, LogHandler &log_handler) - : bp_manager_(bp_manager), frame_manager_(frame_manager), dblwr_manager_(dblwr_manager), log_handler_(*this, log_handler) +DiskBufferPool::DiskBufferPool(BufferPoolManager &bp_manager, BPFrameManager &frame_manager, + DoubleWriteBuffer &dblwr_manager, LogHandler &log_handler) + : bp_manager_(bp_manager), + frame_manager_(frame_manager), + dblwr_manager_(dblwr_manager), + log_handler_(*this, log_handler) {} DiskBufferPool::~DiskBufferPool() @@ -237,7 +240,7 @@ RC DiskBufferPool::open_file(const char *file_name) file_desc_ = fd; Page header_page; - int ret = readn(file_desc_, &header_page, sizeof(header_page)); + int ret = readn(file_desc_, &header_page, sizeof(header_page)); if (ret != 0) { LOG_ERROR("Failed to read first page of %s, due to %s.", file_name, strerror(errno)); close(fd); @@ -246,7 +249,7 @@ RC DiskBufferPool::open_file(const char *file_name) } BPFileHeader *tmp_file_header = reinterpret_cast(header_page.data); - buffer_pool_id_ = tmp_file_header->buffer_pool_id; + buffer_pool_id_ = tmp_file_header->buffer_pool_id; RC rc = allocate_frame(BP_HEADER_PAGE, &hdr_frame_); if (rc != RC::SUCCESS) { @@ -364,7 +367,7 @@ RC DiskBufferPool::allocate_page(Frame **frame) // TODO, do we need clean the loaded page's data? hdr_frame_->mark_dirty(); LSN lsn = 0; - rc = log_handler_.allocate_page(i, lsn); + rc = log_handler_.allocate_page(i, lsn); if (OB_FAIL(rc)) { LOG_ERROR("Failed to log allocate page %d, rc=%s", i, strrc(rc)); // 忽略了错误 @@ -386,7 +389,7 @@ RC DiskBufferPool::allocate_page(Frame **frame) } LSN lsn = 0; - rc = log_handler_.allocate_page(file_header_->page_count, lsn); + rc = log_handler_.allocate_page(file_header_->page_count, lsn); if (OB_FAIL(rc)) { LOG_ERROR("Failed to log allocate page %d, rc=%s", file_header_->page_count, strrc(rc)); // 忽略了错误 @@ -436,9 +439,9 @@ RC DiskBufferPool::dispose_page(PageNum page_num) LOG_ERROR("Failed to dispose page %d, because it is the first page. filename=%s", page_num, file_name_.c_str()); return RC::INTERNAL; } - + scoped_lock lock_guard(lock_); - Frame *used_frame = frame_manager_.get(id(), page_num); + Frame *used_frame = frame_manager_.get(id(), page_num); if (used_frame != nullptr) { ASSERT("the page try to dispose is in use. frame:%s", used_frame->to_string().c_str()); frame_manager_.free(id(), page_num, used_frame); @@ -447,7 +450,7 @@ RC DiskBufferPool::dispose_page(PageNum page_num) } LSN lsn = 0; - RC rc = log_handler_.deallocate_page(page_num, lsn); + RC rc = log_handler_.deallocate_page(page_num, lsn); if (OB_FAIL(rc)) { LOG_ERROR("Failed to log deallocate page %d, rc=%s", page_num, strrc(rc)); // ignore error handle @@ -492,7 +495,7 @@ RC DiskBufferPool::purge_page(PageNum page_num) { scoped_lock lock_guard(lock_); - Frame *used_frame = frame_manager_.get(id(), page_num); + Frame *used_frame = frame_manager_.get(id(), page_num); if (used_frame != nullptr) { return purge_frame(page_num, used_frame); } @@ -647,7 +650,7 @@ RC DiskBufferPool::redo_allocate_page(LSN lsn, PageNum page_num) file_header_->page_count++; hdr_frame_->set_lsn(lsn); hdr_frame_->mark_dirty(); - + // TODO 应该检查文件是否足够大，包含了当前新分配的页面 Bitmap bitmap(file_header_->bitmap, file_header_->page_count); @@ -731,13 +734,13 @@ RC DiskBufferPool::check_page_num(PageNum page_num) RC DiskBufferPool::load_page(PageNum page_num, Frame *frame) { Page &page = frame->page(); - RC rc = dblwr_manager_.read_page(this, page_num, page); + RC rc = dblwr_manager_.read_page(this, page_num, page); if (OB_SUCC(rc)) { return rc; } scoped_lock lock_guard(wr_lock_); - int64_t offset = ((int64_t)page_num) * BP_PAGE_SIZE; + int64_t offset = ((int64_t)page_num) * BP_PAGE_SIZE; if (lseek(file_desc_, offset, SEEK_SET) == -1) { LOG_ERROR("Failed to load page %s:%d, due to failed to lseek:%s.", file_name_.c_str(), page_num, strerror(errno)); @@ -891,7 +894,7 @@ RC BufferPoolManager::flush_page(Frame &frame) int buffer_pool_id = frame.buffer_pool_id(); scoped_lock lock_guard(lock_); - auto iter = id_to_buffer_pools_.find(buffer_pool_id); + auto iter = id_to_buffer_pools_.find(buffer_pool_id); if (iter == id_to_buffer_pools_.end()) { LOG_WARN("unknown buffer pool of id %d", buffer_pool_id); return RC::INTERNAL; @@ -912,8 +915,7 @@ RC BufferPoolManager::get_buffer_pool(int32_t id, DiskBufferPool *&bp) LOG_WARN("unknown buffer pool of id %d", id); return RC::INTERNAL; } - + bp = iter->second; return RC::SUCCESS; } - diff --git a/src/observer/storage/buffer/disk_buffer_pool.h b/src/observer/storage/buffer/disk_buffer_pool.h index 9ff02922..5f930e2e 100644 --- a/src/observer/storage/buffer/disk_buffer_pool.h +++ b/src/observer/storage/buffer/disk_buffer_pool.h @@ -351,5 +351,5 @@ class BufferPoolManager final common::Mutex lock_; unordered_map buffer_pools_; unordered_map id_to_buffer_pools_; - atomic next_buffer_pool_id_{1}; // 系统启动时，会打开所有的表，这样就可以知道当前系统最大的ID是多少了 + atomic next_buffer_pool_id_{1}; // 系统启动时，会打开所有的表，这样就可以知道当前系统最大的ID是多少了 }; diff --git a/src/observer/storage/buffer/double_write_buffer.cpp b/src/observer/storage/buffer/double_write_buffer.cpp index 679c4389..f1120455 100644 --- a/src/observer/storage/buffer/double_write_buffer.cpp +++ b/src/observer/storage/buffer/double_write_buffer.cpp @@ -32,25 +32,24 @@ struct DoubleWritePage public: DoubleWritePageKey key; - int32_t page_index = -1; /// 页面在double write buffer文件中的页索引 - bool valid = true; /// 表示页面是否有效，在页面被删除时，需要同时标记磁盘上的值。 - Page page; + int32_t page_index = -1; /// 页面在double write buffer文件中的页索引 + bool valid = true; /// 表示页面是否有效，在页面被删除时，需要同时标记磁盘上的值。 + Page page; static const int32_t SIZE; }; DoubleWritePage::DoubleWritePage(int32_t buffer_pool_id, PageNum page_num, int32_t page_index, Page &_page) - : key{buffer_pool_id, page_num}, page_index(page_index), page(_page) + : key{buffer_pool_id, page_num}, page_index(page_index), page(_page) {} const int32_t DoubleWritePage::SIZE = sizeof(DoubleWritePage); const int32_t DoubleWriteBufferHeader::SIZE = sizeof(DoubleWriteBufferHeader); -DiskDoubleWriteBuffer::DiskDoubleWriteBuffer(BufferPoolManager &bp_manager, int max_pages /*=16*/) - : max_pages_(max_pages), bp_manager_(bp_manager) -{ -} +DiskDoubleWriteBuffer::DiskDoubleWriteBuffer(BufferPoolManager &bp_manager, int max_pages /*=16*/) + : max_pages_(max_pages), bp_manager_(bp_manager) +{} DiskDoubleWriteBuffer::~DiskDoubleWriteBuffer() { @@ -64,7 +63,7 @@ RC DiskDoubleWriteBuffer::open_file(const char *filename) LOG_ERROR("Double write buffer has already opened. file desc=%d", file_desc_); return RC::BUFFERPOOL_OPEN; } - + int fd = open(filename, O_CREAT | O_RDWR, 0644); if (fd < 0) { LOG_ERROR("Failed to open or creat %s, due to %s.", filename, strerror(errno)); @@ -97,9 +96,9 @@ RC DiskDoubleWriteBuffer::flush_page() RC DiskDoubleWriteBuffer::add_page(DiskBufferPool *bp, PageNum page_num, Page &page) { - scoped_lock lock_guard(lock_); + scoped_lock lock_guard(lock_); DoubleWritePageKey key{bp->id(), page_num}; - auto iter = dblwr_pages_.find(key); + auto iter = dblwr_pages_.find(key); if (iter != dblwr_pages_.end()) { iter->second->page = page; LOG_TRACE("[cache hit]add page into double write buffer. buffer_pool_id:%d,page_num:%d,lsn=%d, dwb size=%d", @@ -147,7 +146,7 @@ RC DiskDoubleWriteBuffer::add_page(DiskBufferPool *bp, PageNum page_num, Page &p RC DiskDoubleWriteBuffer::write_page_internal(DoubleWritePage *page) { int32_t page_index = page->page_index; - int64_t offset = page_index * DoubleWritePage::SIZE + DoubleWriteBufferHeader::SIZE; + int64_t offset = page_index * DoubleWritePage::SIZE + DoubleWriteBufferHeader::SIZE; if (lseek(file_desc_, offset, SEEK_SET) == -1) { LOG_ERROR("Failed to add page %lld of %d due to failed to seek %s.", offset, file_desc_, strerror(errno)); return RC::IOERR_SEEK; @@ -181,9 +180,9 @@ RC DiskDoubleWriteBuffer::write_page(DoubleWritePage *dblwr_page) RC DiskDoubleWriteBuffer::read_page(DiskBufferPool *bp, PageNum page_num, Page &page) { - scoped_lock lock_guard(lock_); + scoped_lock lock_guard(lock_); DoubleWritePageKey key{bp->id(), page_num}; - auto iter = dblwr_pages_.find(key); + auto iter = dblwr_pages_.find(key); if (iter != dblwr_pages_.end()) { page = iter->second->page; LOG_TRACE("double write buffer read page success. bp id=%d, page_num:%d, lsn:%d", bp->id(), page_num, page.lsn); @@ -196,7 +195,7 @@ RC DiskDoubleWriteBuffer::read_page(DiskBufferPool *bp, PageNum page_num, Page & RC DiskDoubleWriteBuffer::clear_pages(DiskBufferPool *buffer_pool) { vector spec_pages; - + auto remove_pred = [&spec_pages, buffer_pool](const pair &pair) { DoubleWritePage *dbl_page = pair.second; if (buffer_pool->id() == dbl_page->key.buffer_pool_id) { @@ -267,9 +266,9 @@ RC DiskDoubleWriteBuffer::load_pages() return RC::IOERR_SEEK; } - auto dblwr_page = make_unique(); - Page &page = dblwr_page->page; - page.check_sum = (CheckSum)-1; + auto dblwr_page = make_unique(); + Page &page = dblwr_page->page; + page.check_sum = (CheckSum)-1; ret = readn(file_desc_, dblwr_page.get(), DoubleWritePage::SIZE); if (ret != 0) { @@ -291,14 +290,10 @@ RC DiskDoubleWriteBuffer::load_pages() return RC::SUCCESS; } -RC DiskDoubleWriteBuffer::recover() -{ - return flush_page(); -} +RC DiskDoubleWriteBuffer::recover() { return flush_page(); } //////////////////////////////////////////////////////////////// RC VacuousDoubleWriteBuffer::add_page(DiskBufferPool *bp, PageNum page_num, Page &page) { return bp->write_page(page_num, page); } - diff --git a/src/observer/storage/buffer/frame.cpp b/src/observer/storage/buffer/frame.cpp index b25ec67c..93e1e3ea 100644 --- a/src/observer/storage/buffer/frame.cpp +++ b/src/observer/storage/buffer/frame.cpp @@ -81,7 +81,13 @@ void Frame::write_latch(intptr_t xid) ++write_recursive_count_; TRACE("frame write lock success." "this=%p, pin=%d, frameId=%s, write locker=%lx(recursive=%d), xid=%lx, lbt=%s", - this, pin_count_.load(), frame_id_.to_string().c_str(), write_locker_, write_recursive_count_, xid, lbt()); + this, + pin_count_.load(), + frame_id_.to_string().c_str(), + write_locker_, + write_recursive_count_, + xid, + lbt()); #endif } @@ -103,7 +109,11 @@ void Frame::write_unlatch(intptr_t xid) write_locker_, this, pin_count_.load(), frame_id_.to_string().c_str(), xid, lbt()); TRACE("frame write unlock success. this=%p, pin=%d, frameId=%s, xid=%lx, lbt=%s", - this, pin_count_.load(), frame_id_.to_string().c_str(), xid, lbt()); + this, + pin_count_.load(), + frame_id_.to_string().c_str(), + xid, + lbt()); if (--write_recursive_count_ == 0) { write_locker_ = 0; @@ -138,7 +148,12 @@ void Frame::read_latch(intptr_t xid) ++read_lockers_[xid]; TRACE("frame read lock success." "this=%p, pin=%d, frameId=%s, xid=%lx, recursive=%d, lbt=%s", - this, pin_count_.load(), frame_id_.to_string().c_str(), xid, read_lockers_[xid], lbt()); + this, + pin_count_.load(), + frame_id_.to_string().c_str(), + xid, + read_lockers_[xid], + lbt()); #endif } } @@ -166,7 +181,12 @@ bool Frame::try_read_latch() ++read_lockers_[xid]; TRACE("frame read lock success." "this=%p, pin=%d, frameId=%s, xid=%lx, recursive=%d, lbt=%s", - this, pin_count_.load(), frame_id_.to_string().c_str(), xid, read_lockers_[xid], lbt()); + this, + pin_count_.load(), + frame_id_.to_string().c_str(), + xid, + read_lockers_[xid], + lbt()); debug_lock_.unlock(); #endif } @@ -203,7 +223,11 @@ void Frame::read_unlatch(intptr_t xid) TRACE("frame read unlock success." "this=%p, pin=%d, frameId=%s, xid=%lx, lbt=%s", - this, pin_count_.load(), frame_id_.to_string().c_str(), xid, lbt()); + this, + pin_count_.load(), + frame_id_.to_string().c_str(), + xid, + lbt()); lock_.unlock_shared(); } @@ -217,8 +241,13 @@ void Frame::pin() TRACE("after frame pin. " "this=%p, write locker=%lx, read locker has xid %d? pin=%d, frameId=%s, xid=%lx, lbt=%s", - this, write_locker_, read_lockers_.find(xid) != read_lockers_.end(), - pin_count, frame_id_.to_string().c_str(), xid, lbt()); + this, + write_locker_, + read_lockers_.find(xid) != read_lockers_.end(), + pin_count, + frame_id_.to_string().c_str(), + xid, + lbt()); } int Frame::unpin() @@ -235,17 +264,14 @@ int Frame::unpin() int pin_count = --pin_count_; TRACE("after frame unpin. " "this=%p, write locker=%lx, read locker has xid? %d, pin=%d, frameId=%s, xid=%lx, lbt=%s", - this, write_locker_, read_lockers_.find(xid) != read_lockers_.end(), - pin_count, frame_id_.to_string().c_str(), xid, lbt()); - - if (0 == pin_count) { - ASSERT(write_locker_ == 0, - "frame unpin to 0 failed while someone hold the write lock. write locker=%lx, frameId=%s, xid=%lx", - write_locker_, frame_id_.to_string().c_str(), xid); - ASSERT(read_lockers_.empty(), - "frame unpin to 0 failed while someone hold the read locks. reader num=%d, frameId=%s, xid=%lx", - read_lockers_.size(), frame_id_.to_string().c_str(), xid); - } + this, + write_locker_, + read_lockers_.find(xid) != read_lockers_.end(), + pin_count, + frame_id_.to_string().c_str(), + xid, + lbt()); + return pin_count; } @@ -261,7 +287,6 @@ void Frame::access() { acc_time_ = current_time(); } string Frame::to_string() const { stringstream ss; - ss << "frame id:" << frame_id().to_string() << ", dirty=" << dirty() << ", pin=" << pin_count() - << ", lsn=" << lsn(); + ss << "frame id:" << frame_id().to_string() << ", dirty=" << dirty() << ", pin=" << pin_count() << ", lsn=" << lsn(); return ss.str(); } diff --git a/src/observer/storage/buffer/page.h b/src/observer/storage/buffer/page.h index d180aec0..8b1fd94b 100644 --- a/src/observer/storage/buffer/page.h +++ b/src/observer/storage/buffer/page.h @@ -23,7 +23,7 @@ static constexpr PageNum BP_INVALID_PAGE_NUM = -1; static constexpr PageNum BP_HEADER_PAGE = 0; -static constexpr const int BP_PAGE_SIZE = (1 << 13); +static constexpr const int BP_PAGE_SIZE = (1 << 17); // 128KB static constexpr const int BP_PAGE_DATA_SIZE = (BP_PAGE_SIZE - sizeof(PageNum) - sizeof(LSN) - sizeof(CheckSum)); /** diff --git a/src/observer/storage/clog/disk_log_handler.cpp b/src/observer/storage/clog/disk_log_handler.cpp index fdce4672..fcaebe10 100644 --- a/src/observer/storage/clog/disk_log_handler.cpp +++ b/src/observer/storage/clog/disk_log_handler.cpp @@ -75,7 +75,7 @@ RC DiskLogHandler::await_termination() RC DiskLogHandler::replay(LogReplayer &replayer, LSN start_lsn) { - LSN max_lsn = 0; + LSN max_lsn = 0; auto replay_callback = [&replayer, &max_lsn](LogEntry &entry) -> RC { if (entry.lsn() > max_lsn) { max_lsn = entry.lsn(); @@ -99,10 +99,10 @@ RC DiskLogHandler::replay(LogReplayer &replayer, LSN start_lsn) return rc; } -RC DiskLogHandler::iterate(function consumer, LSN start_lsn) +RC DiskLogHandler::iterate(function consumer, LSN start_lsn) { vector log_files; - RC rc = file_manager_.list_files(log_files, start_lsn); + RC rc = file_manager_.list_files(log_files, start_lsn); if (OB_FAIL(rc)) { LOG_WARN("failed to list clog files. rc=%s", strrc(rc)); return rc; @@ -172,7 +172,7 @@ void DiskLogHandler::thread_func() LOG_INFO("log handler thread started"); LogFileWriter file_writer; - + RC rc = RC::SUCCESS; while (running_.load() || entry_buffer_.entry_number() > 0) { if (!file_writer.valid() || rc == RC::LOG_FILE_FULL) { @@ -193,7 +193,7 @@ void DiskLogHandler::thread_func() } int flush_count = 0; - rc = entry_buffer_.flush(file_writer, flush_count); + rc = entry_buffer_.flush(file_writer, flush_count); if (OB_FAIL(rc) && RC::LOG_FILE_FULL != rc) { LOG_WARN("failed to flush log entry buffer. rc=%s", strrc(rc)); } diff --git a/src/observer/storage/clog/log_buffer.cpp b/src/observer/storage/clog/log_buffer.cpp index 4cbb1060..11fce081 100644 --- a/src/observer/storage/clog/log_buffer.cpp +++ b/src/observer/storage/clog/log_buffer.cpp @@ -44,7 +44,7 @@ RC LogEntryBuffer::append(LSN &lsn, LogModule module, vector &&data) } LogEntry entry; - RC rc = entry.init(lsn, module, std::move(data)); + RC rc = entry.init(lsn, module, std::move(data)); if (OB_FAIL(rc)) { LOG_WARN("failed to init log entry. rc=%s", strrc(rc)); return rc; @@ -78,7 +78,7 @@ RC LogEntryBuffer::flush(LogFileWriter &writer, int &count) entries_.pop_front(); bytes_ -= entry.total_size(); } - + RC rc = writer.write(entry); if (OB_FAIL(rc)) { lock_guard guard(mutex_); @@ -91,16 +91,10 @@ RC LogEntryBuffer::flush(LogFileWriter &writer, int &count) flushed_lsn_ = entry.lsn(); } } - + return RC::SUCCESS; } -int64_t LogEntryBuffer::bytes() const -{ - return bytes_.load(); -} +int64_t LogEntryBuffer::bytes() const { return bytes_.load(); } -int32_t LogEntryBuffer::entry_number() const -{ - return entries_.size(); -} +int32_t LogEntryBuffer::entry_number() const { return entries_.size(); } diff --git a/src/observer/storage/clog/log_buffer.h b/src/observer/storage/clog/log_buffer.h index 68046f72..77cfe575 100644 --- a/src/observer/storage/clog/log_buffer.h +++ b/src/observer/storage/clog/log_buffer.h @@ -68,7 +68,7 @@ class LogEntryBuffer LSN flushed_lsn() const { return flushed_lsn_.load(); } private: - mutex mutex_; /// 当前数据结构一定会在多线程中访问，所以强制使用有效的锁，而不是有条件生效的common::Mutex + mutex mutex_; /// 当前数据结构一定会在多线程中访问，所以强制使用有效的锁，而不是有条件生效的common::Mutex deque entries_; /// 日志缓冲区 atomic bytes_; /// 当前缓冲区中的日志数据大小 diff --git a/src/observer/storage/clog/log_entry.cpp b/src/observer/storage/clog/log_entry.cpp index 3d34dad1..35672a43 100644 --- a/src/observer/storage/clog/log_entry.cpp +++ b/src/observer/storage/clog/log_entry.cpp @@ -24,9 +24,7 @@ const int32_t LogHeader::SIZE = sizeof(LogHeader); string LogHeader::to_string() const { stringstream ss; - ss << "lsn=" << lsn - << ", size=" << size - << ", module_id=" << module_id << ":" << LogModule(module_id).name(); + ss << "lsn=" << lsn << ", size=" << size << ", module_id=" << module_id << ":" << LogModule(module_id).name(); return ss.str(); } @@ -35,16 +33,16 @@ string LogHeader::to_string() const // class LogEntry LogEntry::LogEntry() { - header_.lsn = 0; + header_.lsn = 0; header_.size = 0; } LogEntry::LogEntry(LogEntry &&other) { header_ = other.header_; - data_ = std::move(other.data_); + data_ = std::move(other.data_); - other.header_.lsn = 0; + other.header_.lsn = 0; other.header_.size = 0; } @@ -55,9 +53,9 @@ LogEntry &LogEntry::operator=(LogEntry &&other) } header_ = other.header_; - data_ = std::move(other.data_); + data_ = std::move(other.data_); - other.header_.lsn = 0; + other.header_.lsn = 0; other.header_.size = 0; return *this; @@ -75,14 +73,11 @@ RC LogEntry::init(LSN lsn, LogModule module, vector &&data) return RC::INVALID_ARGUMENT; } - header_.lsn = lsn; + header_.lsn = lsn; header_.module_id = module.index(); - header_.size = static_cast(data.size()); - data_ = std::move(data); + header_.size = static_cast(data.size()); + data_ = std::move(data); return RC::SUCCESS; } -string LogEntry::to_string() const -{ - return header_.to_string(); -} +string LogEntry::to_string() const { return header_.to_string(); } diff --git a/src/observer/storage/clog/log_file.cpp b/src/observer/storage/clog/log_file.cpp index 18bb0e76..88b1caef 100644 --- a/src/observer/storage/clog/log_file.cpp +++ b/src/observer/storage/clog/log_file.cpp @@ -141,10 +141,7 @@ RC LogFileReader::skip_to(LSN start_lsn) } //////////////////////////////////////////////////////////////////////////////// // LogFileWriter -LogFileWriter::~LogFileWriter() -{ - (void)this->close(); -} +LogFileWriter::~LogFileWriter() { (void)this->close(); } RC LogFileWriter::open(const char *filename, int end_lsn) { @@ -153,7 +150,7 @@ RC LogFileWriter::open(const char *filename, int end_lsn) } filename_ = filename; - end_lsn_ = end_lsn; + end_lsn_ = end_lsn; fd_ = ::open(filename, O_WRONLY | O_APPEND | O_CREAT | O_SYNC, 0644); if (fd_ < 0) { @@ -214,27 +211,18 @@ RC LogFileWriter::write(LogEntry &entry) return RC::SUCCESS; } -bool LogFileWriter::valid() const -{ - return fd_ >= 0; -} +bool LogFileWriter::valid() const { return fd_ >= 0; } -bool LogFileWriter::full() const -{ - return last_lsn_ >= end_lsn_; -} +bool LogFileWriter::full() const { return last_lsn_ >= end_lsn_; } -string LogFileWriter::to_string() const -{ - return filename_; -} +string LogFileWriter::to_string() const { return filename_; } //////////////////////////////////////////////////////////////////////////////// // LogFileManager RC LogFileManager::init(const char *directory, int max_entry_number_per_file) { - directory_ = filesystem::absolute(filesystem::path(directory)); + directory_ = filesystem::absolute(filesystem::path(directory)); max_entry_number_per_file_ = max_entry_number_per_file; // 检查目录是否存在，不存在就创建出来 @@ -242,7 +230,7 @@ RC LogFileManager::init(const char *directory, int max_entry_number_per_file) LOG_INFO("directory is not exist. directory=%s", directory_.c_str()); error_code ec; - bool ret = filesystem::create_directories(directory_, ec); + bool ret = filesystem::create_directories(directory_, ec); if (!ret) { LOG_WARN("create directory failed. directory=%s, error=%s", directory_.c_str(), ec.message().c_str()); return RC::FILE_CREATE; @@ -256,8 +244,8 @@ RC LogFileManager::init(const char *directory, int max_entry_number_per_file) } string filename = dir_entry.path().filename().string(); - LSN lsn = 0; - RC rc = get_lsn_from_filename(filename, lsn); + LSN lsn = 0; + RC rc = get_lsn_from_filename(filename, lsn); if (OB_FAIL(rc)) { LOG_TRACE("invalid log file name. filename=%s", filename.c_str()); continue; @@ -283,7 +271,8 @@ RC LogFileManager::get_lsn_from_filename(const string &filename, LSN &lsn) return RC::INVALID_ARGUMENT; } - string_view lsn_str(filename.data() + strlen(file_prefix_), filename.length() - strlen(file_suffix_) - strlen(file_prefix_)); + string_view lsn_str( + filename.data() + strlen(file_prefix_), filename.length() - strlen(file_suffix_) - strlen(file_prefix_)); from_chars_result result = from_chars(lsn_str.data(), lsn_str.data() + lsn_str.size(), lsn); if (result.ec != errc()) { LOG_TRACE("invalid log file name: cannot calc lsn. filename=%s, error=%s", @@ -318,8 +307,7 @@ RC LogFileManager::last_file(LogFileWriter &file_writer) file_writer.close(); auto last_file_item = log_files_.rbegin(); - return file_writer.open(last_file_item->second.c_str(), - last_file_item->first + max_entry_number_per_file_ - 1); + return file_writer.open(last_file_item->second.c_str(), last_file_item->first + max_entry_number_per_file_ - 1); } RC LogFileManager::next_file(LogFileWriter &file_writer) @@ -331,7 +319,7 @@ RC LogFileManager::next_file(LogFileWriter &file_writer) lsn = log_files_.rbegin()->first + max_entry_number_per_file_; } - string filename = file_prefix_ + std::to_string(lsn) + file_suffix_; + string filename = file_prefix_ + std::to_string(lsn) + file_suffix_; filesystem::path file_path = directory_ / filename; log_files_.emplace(lsn, file_path); diff --git a/src/observer/storage/common/column.cpp b/src/observer/storage/common/column.cpp index 2d4f61dd..6ea43384 100644 --- a/src/observer/storage/common/column.cpp +++ b/src/observer/storage/common/column.cpp @@ -78,12 +78,12 @@ void Column::reset() if (data_ != nullptr && own_) { delete[] data_; } - data_ = nullptr; - count_ = 0; - capacity_ = 0; - own_ = false; - attr_type_ = AttrType::UNDEFINED; - attr_len_ = -1; + data_ = nullptr; + count_ = 0; + capacity_ = 0; + own_ = false; + attr_type_ = AttrType::UNDEFINED; + attr_len_ = -1; } RC Column::append_one(char *data) { return append(data, 1); } diff --git a/src/observer/storage/common/meta_util.cpp b/src/observer/storage/common/meta_util.cpp index 00f4c5ff..136bf8ce 100644 --- a/src/observer/storage/common/meta_util.cpp +++ b/src/observer/storage/common/meta_util.cpp @@ -28,6 +28,11 @@ string table_meta_file(const char *base_dir, const char *table_name) return filesystem::path(base_dir) / (string(table_name) + TABLE_META_SUFFIX); } +string vtable_meta_file(const char *base_dir, const char *table_name) +{ + return filesystem::path(base_dir) / (string(table_name) + VTABLE_META_SUFFIX); +} + string table_data_file(const char *base_dir, const char *table_name) { return filesystem::path(base_dir) / (string(table_name) + TABLE_DATA_SUFFIX); diff --git a/src/observer/storage/common/meta_util.h b/src/observer/storage/common/meta_util.h index 2eaacd61..4ef48a43 100644 --- a/src/observer/storage/common/meta_util.h +++ b/src/observer/storage/common/meta_util.h @@ -17,11 +17,15 @@ See the Mulan PSL v2 for more details. */ static constexpr const char *DB_META_SUFFIX = ".db"; static constexpr const char *TABLE_META_SUFFIX = ".table"; +static constexpr const char *VTABLE_META_SUFFIX = ".vtable"; static constexpr const char *TABLE_META_FILE_PATTERN = ".*\\.table$"; -static constexpr const char *TABLE_DATA_SUFFIX = ".data"; -static constexpr const char *TABLE_INDEX_SUFFIX = ".index"; +static constexpr const char *VTABLE_META_FILE_PATTERN = + ".*\\.vtable$"; // 视图的 meta 命名和基表不一样，用于持久化时做区分 +static constexpr const char *TABLE_DATA_SUFFIX = ".data"; +static constexpr const char *TABLE_INDEX_SUFFIX = ".index"; string db_meta_file(const char *base_dir, const char *db_name); string table_meta_file(const char *base_dir, const char *table_name); +string vtable_meta_file(const char *base_dir, const char *table_name); string table_data_file(const char *base_dir, const char *table_name); string table_index_file(const char *base_dir, const char *table_name, const char *index_name); diff --git a/src/observer/storage/db/db.cpp b/src/observer/storage/db/db.cpp index 74e9a169..a170a9f5 100644 --- a/src/observer/storage/db/db.cpp +++ b/src/observer/storage/db/db.cpp @@ -16,6 +16,7 @@ See the Mulan PSL v2 for more details. */ #include #include +#include #include #include @@ -29,6 +30,7 @@ See the Mulan PSL v2 for more details. */ #include "storage/trx/trx.h" #include "storage/clog/disk_log_handler.h" #include "storage/clog/integrated_log_replayer.h" +#include "storage/table/view.h" using namespace common; @@ -161,16 +163,74 @@ RC Db::create_table(const char *table_name, span attribut return RC::SUCCESS; } -Table *Db::find_table(const char *table_name) const +RC Db::create_table(const char *table_name, std::vector attr_names, std::string select_sql, + SelectStmt *select_stmt, const StorageFormat storage_format) { - unordered_map::const_iterator iter = opened_tables_.find(table_name); + RC rc = RC::SUCCESS; + // check table_name + if (opened_tables_.count(table_name) != 0) { + LOG_WARN("%s has been opened before.", table_name); + return RC::SCHEMA_TABLE_EXIST; + } + + // 文件路径可以移到Table模块 + string table_file_path = vtable_meta_file(path_.c_str(), table_name); + View *table = new View; + int32_t table_id = next_table_id_++; + rc = table->create(this, + table_id, + table_file_path.c_str(), + table_name, + path_.c_str(), + std::move(attr_names), + std::move(select_sql), + select_stmt, + storage_format); + if (rc != RC::SUCCESS) { + LOG_ERROR("Failed to create table %s.", table_name); + delete table; + return rc; + } + + opened_tables_[table_name] = table; + LOG_INFO("Create table success. table name=%s, table_id:%d", table_name, table_id); + return RC::SUCCESS; +} + +RC Db::drop_table(const char *table_name) +{ + // check table_name + auto it = opened_tables_.find(table_name); + if (it == opened_tables_.end()) { + LOG_WARN("%s has not been opened before.", table_name); + return RC::SCHEMA_TABLE_NOT_EXIST; + } + + // drop table + auto table = it->second; + auto rc = table->drop(); + if (rc != RC::SUCCESS) { + LOG_ERROR("Failed to drop table %s.", table_name); + return rc; + } + + opened_tables_.erase(table_name); + // release memory + delete table; + LOG_INFO("Drop table success. table name=%s", table_name); + return RC::SUCCESS; +} + +BaseTable *Db::find_table(const char *table_name) const +{ + unordered_map::const_iterator iter = opened_tables_.find(table_name); if (iter != opened_tables_.end()) { return iter->second; } return nullptr; } -Table *Db::find_table(int32_t table_id) const +BaseTable *Db::find_table(int32_t table_id) const { for (auto pair : opened_tables_) { if (pair.second->table_id() == table_id) { @@ -182,7 +242,8 @@ Table *Db::find_table(int32_t table_id) const RC Db::open_all_tables() { - vector table_meta_files; + std::vector table_meta_files; + std::vector view_meta_files; int ret = list_file(path_.c_str(), TABLE_META_FILE_PATTERN, table_meta_files); if (ret < 0) { @@ -190,10 +251,16 @@ RC Db::open_all_tables() return RC::IOERR_READ; } + ret = list_file(path_.c_str(), VTABLE_META_FILE_PATTERN, view_meta_files); + if (ret < 0) { + LOG_ERROR("Failed to list view meta files under %s.", path_.c_str()); + return RC::IOERR_READ; + } + RC rc = RC::SUCCESS; for (const string &filename : table_meta_files) { - Table *table = new Table(); - rc = table->open(this, filename.c_str(), path_.c_str()); + BaseTable *table = new Table(); + rc = table->open(this, filename.c_str(), path_.c_str()); if (rc != RC::SUCCESS) { delete table; LOG_ERROR("Failed to open table. filename=%s", filename.c_str()); @@ -211,10 +278,50 @@ RC Db::open_all_tables() if (table->table_id() >= next_table_id_) { next_table_id_ = table->table_id() + 1; } + opened_tables_[table->name()] = table; LOG_INFO("Open table: %s, file: %s", table->name(), filename.c_str()); } + for (const string &filename : view_meta_files) { + BaseTable *table = new View(); + // 视图的 open 只是初始化了元数据和查询 sql，成员变量未初始化 + rc = table->open(this, filename.c_str(), path_.c_str()); + if (rc != RC::SUCCESS) { + delete table; + LOG_ERROR("Failed to open table. filename=%s", filename.c_str()); + return rc; + } + + if (opened_tables_.count(table->name()) != 0) { + LOG_ERROR("Duplicate table with difference file name. table=%s, the other filename=%s", + table->name(), filename.c_str()); + // 在这里原本先删除table后调用table->name()方法，犯了use-after-free的错误 + delete table; + return RC::INTERNAL; + } + + if (table->table_id() >= next_table_id_) { + next_table_id_ = table->table_id() + 1; + } + + opened_tables_[table->name()] = table; + LOG_INFO("Open table: %s, file: %s", table->name(), filename.c_str()); + } + + // 在所有表都载入后，对于视图要手动调一下 init，初始化成员变量 + for (auto &[_, table] : opened_tables_) { + if (table->type() == TableType::View) { + auto view = dynamic_cast(table); + rc = view->init_member(); + if (OB_FAIL(rc)) { + LOG_ERROR("Failed to initialize view: %s", table->name()); + return rc; + } + LOG_INFO("Successfully initialized view: %s", table->name()); + } + } + LOG_INFO("All table have been opened. num=%d", opened_tables_.size()); return rc; } @@ -233,8 +340,8 @@ RC Db::sync() RC rc = RC::SUCCESS; // 调用所有表的sync函数刷新数据到磁盘 for (const auto &table_pair : opened_tables_) { - Table *table = table_pair.second; - rc = table->sync(); + BaseTable *table = table_pair.second; + rc = table->sync(); if (rc != RC::SUCCESS) { LOG_ERROR("Failed to flush table. table=%s.%s, rc=%d:%s", name_.c_str(), table->name(), rc, strrc(rc)); return rc; diff --git a/src/observer/storage/db/db.h b/src/observer/storage/db/db.h index 8eee8943..9597a51e 100644 --- a/src/observer/storage/db/db.h +++ b/src/observer/storage/db/db.h @@ -24,11 +24,13 @@ See the Mulan PSL v2 for more details. */ #include "storage/buffer/disk_buffer_pool.h" #include "storage/clog/disk_log_handler.h" #include "storage/buffer/double_write_buffer.h" +#include "storage/table/base_table.h" class Table; class LogHandler; class BufferPoolManager; class TrxKit; +class SelectStmt; /** * @brief 一个DB实例负责管理一批表 @@ -64,16 +66,29 @@ class Db * @param storage_format 表的存储格式 */ RC create_table(const char *table_name, span attributes, - const StorageFormat storage_format = StorageFormat::ROW_FORMAT); + StorageFormat storage_format = StorageFormat::ROW_FORMAT); + + /** + * @brief 创建一个视图 + * @param table_name 表名 + * @param attr_names 表的属性 + * @param select_sql 查询 sql + * @param select_stmt 查询 stmt + * @param storage_format 表的存储格式 + */ + RC create_table(const char *table_name, std::vector attr_names, std::string select_sql, + SelectStmt *select_stmt, StorageFormat storage_format); + + RC drop_table(const char *table_name); /** * @brief 根据表名查找表 */ - Table *find_table(const char *table_name) const; + BaseTable *find_table(const char *table_name) const; /** * @brief 根据表ID查找表 */ - Table *find_table(int32_t table_id) const; + BaseTable *find_table(int32_t table_id) const; /// @brief 当前数据库的名称 const char *name() const; @@ -111,12 +126,12 @@ class Db RC init_dblwr_buffer(); private: - string name_; ///< 数据库名称 - string path_; ///< 数据库文件存放的目录 - unordered_map opened_tables_; ///< 当前所有打开的表 - unique_ptr buffer_pool_manager_; ///< 当前数据库的buffer pool管理器 - unique_ptr log_handler_; ///< 当前数据库的日志处理器 - unique_ptr trx_kit_; ///< 当前数据库的事务管理器 + string name_; ///< 数据库名称 + string path_; ///< 数据库文件存放的目录 + unordered_map opened_tables_; ///< 当前所有打开的表 + unique_ptr buffer_pool_manager_; ///< 当前数据库的buffer pool管理器 + unique_ptr log_handler_; ///< 当前数据库的日志处理器 + unique_ptr trx_kit_; ///< 当前数据库的事务管理器 /// 给每个table都分配一个ID，用来记录日志。这里假设所有的DDL都不会并发操作，所以相关的数据都不上锁 int32_t next_table_id_ = 0; diff --git a/src/observer/storage/default/default_handler.cpp b/src/observer/storage/default/default_handler.cpp index e310aa91..f860c2a0 100644 --- a/src/observer/storage/default/default_handler.cpp +++ b/src/observer/storage/default/default_handler.cpp @@ -27,7 +27,7 @@ See the Mulan PSL v2 for more details. */ #include "storage/table/table.h" #include "storage/trx/trx.h" -using namespace std; +// using namespace std; DefaultHandler::DefaultHandler() {} @@ -44,9 +44,9 @@ RC DefaultHandler::init(const char *base_dir, const char *trx_kit_name, const ch return RC::INTERNAL; } - base_dir_ = base_dir; - db_dir_ = db_dir; - trx_kit_name_ = trx_kit_name; + base_dir_ = base_dir; + db_dir_ = db_dir; + trx_kit_name_ = trx_kit_name; log_handler_name_ = log_handler_name; const char *sys_db = "sys"; @@ -102,7 +102,27 @@ RC DefaultHandler::create_db(const char *dbname) return RC::SUCCESS; } -RC DefaultHandler::drop_db(const char *dbname) { return RC::INTERNAL; } +RC DefaultHandler::drop_db(const char *dbname) +{ + if (dbname == nullptr || common::is_blank(dbname)) { + LOG_WARN("Invalid db name"); + return RC::INVALID_ARGUMENT; + } + + // 如果对应目录不存在，返回错误 + filesystem::path dbpath = db_dir_ / dbname; + if (!filesystem::is_directory(dbpath)) { + LOG_WARN("Db not exists: %s", dbname); + return RC::SCHEMA_DB_NOT_EXIST; + } + + error_code ec; + if (!filesystem::remove(dbpath, ec)) { + LOG_ERROR("Drop db fail: %s. error=%s", dbpath.c_str(), strerror(errno)); + return RC::IOERR_WRITE; + } + return RC::SUCCESS; +} RC DefaultHandler::open_db(const char *dbname) { @@ -154,7 +174,7 @@ Db *DefaultHandler::find_db(const char *dbname) const return iter->second; } -Table *DefaultHandler::find_table(const char *dbname, const char *table_name) const +BaseTable *DefaultHandler::find_table(const char *dbname, const char *table_name) const { if (dbname == nullptr || table_name == nullptr) { LOG_WARN("Invalid argument. dbname=%p, table_name=%p", dbname, table_name); diff --git a/src/observer/storage/default/default_handler.h b/src/observer/storage/default/default_handler.h index 56d7ae06..d07ef001 100644 --- a/src/observer/storage/default/default_handler.h +++ b/src/observer/storage/default/default_handler.h @@ -85,8 +85,8 @@ class DefaultHandler RC drop_table(const char *dbname, const char *relation_name); public: - Db *find_db(const char *dbname) const; - Table *find_table(const char *dbname, const char *table_name) const; + Db *find_db(const char *dbname) const; + BaseTable *find_table(const char *dbname, const char *table_name) const; RC sync(); diff --git a/src/observer/storage/field/field.h b/src/observer/storage/field/field.h index 7022eb5f..d031c7e2 100644 --- a/src/observer/storage/field/field.h +++ b/src/observer/storage/field/field.h @@ -25,10 +25,10 @@ class Field { public: Field() = default; - Field(const Table *table, const FieldMeta *field) : table_(table), field_(field) {} + Field(const BaseTable *table, const FieldMeta *field) : table_(table), field_(field) {} Field(const Field &) = default; - const Table *table() const { return table_; } + const BaseTable *table() const { return table_; } const FieldMeta *meta() const { return field_; } AttrType attr_type() const { return field_->type(); } @@ -36,7 +36,7 @@ class Field const char *table_name() const { return table_->name(); } const char *field_name() const { return field_->name(); } - void set_table(const Table *table) { this->table_ = table; } + void set_table(const BaseTable *table) { this->table_ = table; } void set_field(const FieldMeta *field) { this->field_ = field; } void set_int(Record &record, int value); @@ -44,7 +44,13 @@ class Field const char *get_data(const Record &record); + bool operator==(const Field &other) const + { + return std::string(table_name()) == std::string(other.table_name()) && + std::string(field_name()) == std::string(other.field_name()); + } + private: - const Table *table_ = nullptr; + const BaseTable *table_ = nullptr; const FieldMeta *field_ = nullptr; }; diff --git a/src/observer/storage/field/field_meta.cpp b/src/observer/storage/field/field_meta.cpp index 9be3bf2e..0f85b62d 100644 --- a/src/observer/storage/field/field_meta.cpp +++ b/src/observer/storage/field/field_meta.cpp @@ -25,16 +25,21 @@ const static Json::StaticString FIELD_OFFSET("offset"); const static Json::StaticString FIELD_LEN("len"); const static Json::StaticString FIELD_VISIBLE("visible"); const static Json::StaticString FIELD_FIELD_ID("FIELD_id"); +const static Json::StaticString FIELD_NULLABLE("nullable"); +const static Json::StaticString FIELD_MUTABLE("mutable"); -FieldMeta::FieldMeta() : attr_type_(AttrType::UNDEFINED), attr_offset_(-1), attr_len_(0), visible_(false), field_id_(0) {} +FieldMeta::FieldMeta() : attr_type_(AttrType::UNDEFINED), attr_offset_(-1), attr_len_(0), visible_(false), field_id_(0) +{} -FieldMeta::FieldMeta(const char *name, AttrType attr_type, int attr_offset, int attr_len, bool visible, int field_id) +FieldMeta::FieldMeta(const char *name, AttrType attr_type, int attr_offset, int attr_len, bool visible, int field_id, + bool nullable, bool is_mutable) { - [[maybe_unused]] RC rc = this->init(name, attr_type, attr_offset, attr_len, visible, field_id); + [[maybe_unused]] RC rc = this->init(name, attr_type, attr_offset, attr_len, visible, field_id, nullable, is_mutable); ASSERT(rc == RC::SUCCESS, "failed to init field meta. rc=%s", strrc(rc)); } -RC FieldMeta::init(const char *name, AttrType attr_type, int attr_offset, int attr_len, bool visible, int field_id) +RC FieldMeta::init(const char *name, AttrType attr_type, int attr_offset, int attr_len, bool visible, int field_id, + bool nullable, bool is_mutable) { if (common::is_blank(name)) { LOG_WARN("Name cannot be empty"); @@ -52,7 +57,9 @@ RC FieldMeta::init(const char *name, AttrType attr_type, int attr_offset, int at attr_len_ = attr_len; attr_offset_ = attr_offset; visible_ = visible; - field_id_ = field_id; + field_id_ = field_id; + nullable_ = nullable; + mutable_ = is_mutable; LOG_INFO("Init a field with name=%s", name); return RC::SUCCESS; @@ -70,20 +77,27 @@ bool FieldMeta::visible() const { return visible_; } int FieldMeta::field_id() const { return field_id_; } +bool FieldMeta::nullable() const { return nullable_; } + +bool FieldMeta::is_mutable() const { return mutable_; } + void FieldMeta::desc(std::ostream &os) const { os << "field name=" << name_ << ", type=" << attr_type_to_string(attr_type_) << ", len=" << attr_len_ - << ", visible=" << (visible_ ? "yes" : "no"); + << ", visible=" << (visible_ ? "yes" : "no") << ", nullable=" << (nullable_ ? "yes" : "no") + << ", mutable=" << (mutable_ ? "yes" : "no"); } void FieldMeta::to_json(Json::Value &json_value) const { - json_value[FIELD_NAME] = name_; - json_value[FIELD_TYPE] = attr_type_to_string(attr_type_); - json_value[FIELD_OFFSET] = attr_offset_; - json_value[FIELD_LEN] = attr_len_; - json_value[FIELD_VISIBLE] = visible_; + json_value[FIELD_NAME] = name_; + json_value[FIELD_TYPE] = attr_type_to_string(attr_type_); + json_value[FIELD_OFFSET] = attr_offset_; + json_value[FIELD_LEN] = attr_len_; + json_value[FIELD_VISIBLE] = visible_; json_value[FIELD_FIELD_ID] = field_id_; + json_value[FIELD_NULLABLE] = nullable_; + json_value[FIELD_MUTABLE] = mutable_; } RC FieldMeta::from_json(const Json::Value &json_value, FieldMeta &field) @@ -93,12 +107,14 @@ RC FieldMeta::from_json(const Json::Value &json_value, FieldMeta &field) return RC::INTERNAL; } - const Json::Value &name_value = json_value[FIELD_NAME]; - const Json::Value &type_value = json_value[FIELD_TYPE]; - const Json::Value &offset_value = json_value[FIELD_OFFSET]; - const Json::Value &len_value = json_value[FIELD_LEN]; - const Json::Value &visible_value = json_value[FIELD_VISIBLE]; + const Json::Value &name_value = json_value[FIELD_NAME]; + const Json::Value &type_value = json_value[FIELD_TYPE]; + const Json::Value &offset_value = json_value[FIELD_OFFSET]; + const Json::Value &len_value = json_value[FIELD_LEN]; + const Json::Value &visible_value = json_value[FIELD_VISIBLE]; const Json::Value &field_id_value = json_value[FIELD_FIELD_ID]; + const Json::Value &nullable_value = json_value[FIELD_NULLABLE]; + const Json::Value &mutable_value = json_value[FIELD_MUTABLE]; if (!name_value.isString()) { LOG_ERROR("Field name is not a string. json value=%s", name_value.toStyledString().c_str()); @@ -126,16 +142,28 @@ RC FieldMeta::from_json(const Json::Value &json_value, FieldMeta &field) return RC::INTERNAL; } + if (!nullable_value.isBool()) { + LOG_ERROR("Nullable id is not a bool value. json value=%s", nullable_value.toStyledString().c_str()); + return RC::INTERNAL; + } + + if (!mutable_value.isBool()) { + LOG_ERROR("Mutable id is not a bool value. json value=%s", mutable_value.toStyledString().c_str()); + return RC::INTERNAL; + } + AttrType type = attr_type_from_string(type_value.asCString()); if (AttrType::UNDEFINED == type) { LOG_ERROR("Got invalid field type. type=%d", type); return RC::INTERNAL; } - const char *name = name_value.asCString(); - int offset = offset_value.asInt(); - int len = len_value.asInt(); - bool visible = visible_value.asBool(); - int field_id = field_id_value.asInt(); - return field.init(name, type, offset, len, visible, field_id); + const char *name = name_value.asCString(); + int offset = offset_value.asInt(); + int len = len_value.asInt(); + bool visible = visible_value.asBool(); + int field_id = field_id_value.asInt(); + bool nullable = nullable_value.asBool(); + bool is_mutable = mutable_value.asBool(); + return field.init(name, type, offset, len, visible, field_id, nullable, is_mutable); } diff --git a/src/observer/storage/field/field_meta.h b/src/observer/storage/field/field_meta.h index 03e9c797..8ccb6a74 100644 --- a/src/observer/storage/field/field_meta.h +++ b/src/observer/storage/field/field_meta.h @@ -30,10 +30,15 @@ class FieldMeta { public: FieldMeta(); - FieldMeta(const char *name, AttrType attr_type, int attr_offset, int attr_len, bool visible, int field_id); + + // 除非字段定义为 not null，否则默认是 nullable 的 + FieldMeta(const char *name, AttrType attr_type, int attr_offset, int attr_len, bool visible, int field_id, + bool nullable = true, bool mutable_ = true); + ~FieldMeta() = default; - RC init(const char *name, AttrType attr_type, int attr_offset, int attr_len, bool visible, int field_id); + RC init(const char *name, AttrType attr_type, int attr_offset, int attr_len, bool visible, int field_id, + bool nullable = true, bool mutable_ = true); public: const char *name() const; @@ -42,6 +47,8 @@ class FieldMeta int len() const; bool visible() const; int field_id() const; + bool nullable() const; + bool is_mutable() const; public: void desc(ostream &os) const; @@ -57,4 +64,6 @@ class FieldMeta int attr_len_; bool visible_; int field_id_; + bool nullable_ = true; + bool mutable_ = true; // 如果是视图的表达式字段，实际上是不可插入或更新的 }; diff --git a/src/observer/storage/index/bplus_tree.cpp b/src/observer/storage/index/bplus_tree.cpp index 7c7a1103..3ed556e3 100644 --- a/src/observer/storage/index/bplus_tree.cpp +++ b/src/observer/storage/index/bplus_tree.cpp @@ -79,15 +79,12 @@ int IndexNodeHandler::min_size() const return max - max / 2; } -void IndexNodeHandler::increase_size(int n) -{ - node_->key_num += n; -} +void IndexNodeHandler::increase_size(int n) { node_->key_num += n; } PageNum IndexNodeHandler::parent_page_num() const { return node_->parent; } -RC IndexNodeHandler::set_parent_page_num(PageNum page_num) -{ +RC IndexNodeHandler::set_parent_page_num(PageNum page_num) +{ RC rc = mtr_.logger().set_parent_page(*this, page_num, this->node_->parent); if (OB_FAIL(rc)) { LOG_WARN("failed to log set parent page. rc=%s", strrc(rc)); @@ -194,20 +191,20 @@ RC LeafIndexNodeHandler::init_empty() LOG_WARN("failed to log init empty leaf node. rc=%s", strrc(rc)); return rc; } - IndexNodeHandler::init_empty(true/*leaf*/); + IndexNodeHandler::init_empty(true /*leaf*/); leaf_node_->next_brother = BP_INVALID_PAGE_NUM; return RC::SUCCESS; } -RC LeafIndexNodeHandler::set_next_page(PageNum page_num) -{ +RC LeafIndexNodeHandler::set_next_page(PageNum page_num) +{ RC rc = mtr_.logger().leaf_set_next_page(*this, page_num, leaf_node_->next_brother); if (OB_FAIL(rc)) { LOG_WARN("failed to log set next page. rc=%s", strrc(rc)); return rc; } - leaf_node_->next_brother = page_num; + leaf_node_->next_brother = page_num; return RC::SUCCESS; } @@ -277,13 +274,14 @@ int LeafIndexNodeHandler::remove(const char *key, const KeyComparator &comparato RC LeafIndexNodeHandler::move_half_to(LeafIndexNodeHandler &other) { - const int size = this->size(); - const int move_index = size / 2; + const int size = this->size(); + const int move_index = size / 2; const int move_item_num = size - move_index; other.append(__item_at(move_index), move_item_num); - RC rc = mtr_.logger().node_remove_items(*this, move_index, span(__item_at(move_index), move_item_num * item_size()), move_item_num); + RC rc = mtr_.logger().node_remove_items( + *this, move_index, span(__item_at(move_index), move_item_num * item_size()), move_item_num); if (OB_FAIL(rc)) { LOG_WARN("failed to log shrink leaf node. rc=%s", strrc(rc)); return rc; @@ -314,7 +312,8 @@ RC LeafIndexNodeHandler::move_to(LeafIndexNodeHandler &other) other.append(__item_at(0), this->size()); other.set_next_page(this->next_page()); - RC rc = mtr_.logger().node_remove_items(*this, 0, span(__item_at(0), this->size() * item_size()), this->size()); + RC rc = mtr_.logger().node_remove_items( + *this, 0, span(__item_at(0), this->size() * item_size()), this->size()); if (OB_FAIL(rc)) { LOG_WARN("failed to log shrink leaf node. rc=%s", strrc(rc)); } @@ -335,15 +334,9 @@ RC LeafIndexNodeHandler::append(const char *items, int num) return recover_insert_items(size(), items, num); } -RC LeafIndexNodeHandler::append(const char *item) -{ - return append(item, 1); -} +RC LeafIndexNodeHandler::append(const char *item) { return append(item, 1); } -RC LeafIndexNodeHandler::preappend(const char *item) -{ - return insert(0, item, item + key_size()); -} +RC LeafIndexNodeHandler::preappend(const char *item) { return insert(0, item, item + key_size()); } char *LeafIndexNodeHandler::__item_at(int index) const { return leaf_node_->array + (index * item_size()); } @@ -381,7 +374,7 @@ bool LeafIndexNodeHandler::validate(const KeyComparator &comparator, DiskBufferP } Frame *parent_frame = nullptr; - RC rc = bp->get_this_page(parent_page_num, &parent_frame); + RC rc = bp->get_this_page(parent_page_num, &parent_frame); if (OB_FAIL(rc)) { LOG_WARN("failed to fetch parent page. page num=%d, rc=%d:%s", parent_page_num, rc, strrc(rc)); return false; @@ -422,7 +415,8 @@ bool LeafIndexNodeHandler::validate(const KeyComparator &comparator, DiskBufferP } ///////////////////////////////////////////////////////////////////////////////// -InternalIndexNodeHandler::InternalIndexNodeHandler(BplusTreeMiniTransaction &mtr, const IndexFileHeader &header, Frame *frame) +InternalIndexNodeHandler::InternalIndexNodeHandler( + BplusTreeMiniTransaction &mtr, const IndexFileHeader &header, Frame *frame) : IndexNodeHandler(mtr, header, frame), internal_node_((InternalIndexNode *)frame->data()) {} @@ -441,13 +435,13 @@ string to_string(const InternalIndexNodeHandler &node, const KeyPrinter &printer return ss.str(); } -RC InternalIndexNodeHandler::init_empty() +RC InternalIndexNodeHandler::init_empty() { RC rc = mtr_.logger().internal_init_empty(*this); if (OB_FAIL(rc)) { LOG_WARN("failed to log init empty internal node. rc=%s", strrc(rc)); } - IndexNodeHandler::init_empty(false/*leaf*/); + IndexNodeHandler::init_empty(false /*leaf*/); return RC::SUCCESS; } RC InternalIndexNodeHandler::create_new_root(PageNum first_page_num, const char *key, PageNum page_num) @@ -502,7 +496,8 @@ RC InternalIndexNodeHandler::move_half_to(InternalIndexNodeHandler &other) return rc; } - mtr_.logger().node_remove_items(*this, move_index, span(__item_at(move_index), move_num * item_size()), move_num); + mtr_.logger().node_remove_items( + *this, move_index, span(__item_at(move_index), move_num * item_size()), move_num); increase_size(-(size - move_index)); return rc; } @@ -550,7 +545,8 @@ void InternalIndexNodeHandler::set_key_at(int index, const char *key) { assert(index >= 0 && index < size()); - mtr_.logger().internal_update_key(*this, index, span(key, key_size()), span(__key_at(index), key_size())); + mtr_.logger().internal_update_key( + *this, index, span(key, key_size()), span(__key_at(index), key_size())); memcpy(__key_at(index), key, key_size()); } @@ -575,7 +571,7 @@ void InternalIndexNodeHandler::remove(int index) assert(index >= 0 && index < size()); BplusTreeLogger &logger = mtr_.logger(); - RC rc = logger.node_remove_items(*this, index, span(__item_at(index), item_size()), 1); + RC rc = logger.node_remove_items(*this, index, span(__item_at(index), item_size()), 1); if (OB_FAIL(rc)) { LOG_WARN("failed to log remove item. rc=%s. node=%s", strrc(rc), to_string(*this).c_str()); } @@ -639,22 +635,22 @@ RC InternalIndexNodeHandler::insert_items(int index, const char *items, int num) recover_insert_items(index, items, num); - LatchMemo &latch_memo = mtr_.latch_memo(); - PageNum this_page_num = this->page_num(); - Frame *frame = nullptr; + LatchMemo &latch_memo = mtr_.latch_memo(); + PageNum this_page_num = this->page_num(); + Frame *frame = nullptr; // 设置所有页面的父页面为当前页面 // 这里会访问大量的页面，可能会将他们从磁盘加载到内存中而占用大量的buffer pool页面 for (int i = 0; i < num; i++) { const PageNum page_num = *(const PageNum *)((items + i * item_size()) + key_size()); - rc = latch_memo.get_page(page_num, frame); + rc = latch_memo.get_page(page_num, frame); if (OB_FAIL(rc)) { LOG_WARN("failed to set child's page num. child page num:%d, this page num=%d, rc=%d:%s", page_num, this_page_num, rc, strrc(rc)); return rc; } IndexNodeHandler child_node(mtr_, header_, frame); - child_node.set_parent_page_num(this_page_num); // 这里虽然对页面做了修改，但是并没有加写锁，因为父页面加了锁 + child_node.set_parent_page_num(this_page_num); // 这里虽然对页面做了修改，但是并没有加写锁，因为父页面加了锁 frame->mark_dirty(); } @@ -664,17 +660,11 @@ RC InternalIndexNodeHandler::insert_items(int index, const char *items, int num) /** * copy items from other node to self's right */ -RC InternalIndexNodeHandler::append(const char *items, int num) -{ - return insert_items(size(), items, num); -} +RC InternalIndexNodeHandler::append(const char *items, int num) { return insert_items(size(), items, num); } RC InternalIndexNodeHandler::append(const char *item) { return this->append(item, 1); } -RC InternalIndexNodeHandler::preappend(const char *item) -{ - return this->insert_items(0, item, 1); -} +RC InternalIndexNodeHandler::preappend(const char *item) { return this->insert_items(0, item, 1); } char *InternalIndexNodeHandler::__item_at(int index) const { return internal_node_->array + (index * item_size()); } @@ -704,7 +694,7 @@ bool InternalIndexNodeHandler::validate(const KeyComparator &comparator, DiskBuf LOG_WARN("this page num=%d, got invalid child page. page num=%d", this->page_num(), page_num); } else { Frame *child_frame = nullptr; - RC rc = bp->get_this_page(page_num, &child_frame); + RC rc = bp->get_this_page(page_num, &child_frame); if (OB_FAIL(rc)) { LOG_WARN("failed to fetch child page while validate internal page. page num=%d, rc=%d:%s", page_num, rc, strrc(rc)); @@ -730,7 +720,7 @@ bool InternalIndexNodeHandler::validate(const KeyComparator &comparator, DiskBuf } Frame *parent_frame = nullptr; - RC rc = bp->get_this_page(parent_page_num, &parent_frame); + RC rc = bp->get_this_page(parent_page_num, &parent_frame); if (OB_FAIL(rc)) { LOG_WARN("failed to fetch parent page. page num=%d, rc=%d:%s", parent_page_num, rc, strrc(rc)); return false; @@ -794,13 +784,8 @@ RC BplusTreeHandler::sync() return disk_buffer_pool_->flush_all_pages(); } -RC BplusTreeHandler::create(LogHandler &log_handler, - BufferPoolManager &bpm, - const char *file_name, - AttrType attr_type, - int attr_length, - int internal_max_size /* = -1*/, - int leaf_max_size /* = -1 */) +RC BplusTreeHandler::create( + LogHandler &log_handler, BufferPoolManager &bpm, const char *file_name, const IndexMeta &index) { RC rc = bpm.create_file(file_name); if (OB_FAIL(rc)) { @@ -818,7 +803,7 @@ RC BplusTreeHandler::create(LogHandler &log_handler, } LOG_INFO("Successfully open index file %s.", file_name); - rc = this->create(log_handler, *bp, attr_type, attr_length, internal_max_size, leaf_max_size); + rc = this->create(log_handler, *bp, index); if (OB_FAIL(rc)) { bpm.close_file(file_name); return rc; @@ -828,20 +813,8 @@ RC BplusTreeHandler::create(LogHandler &log_handler, return rc; } -RC BplusTreeHandler::create(LogHandler &log_handler, - DiskBufferPool &buffer_pool, - AttrType attr_type, - int attr_length, - int internal_max_size /* = -1 */, - int leaf_max_size /* = -1 */) +RC BplusTreeHandler::create(LogHandler &log_handler, DiskBufferPool &buffer_pool, const IndexMeta &index) { - if (internal_max_size < 0) { - internal_max_size = calc_internal_page_capacity(attr_length); - } - if (leaf_max_size < 0) { - leaf_max_size = calc_leaf_page_capacity(attr_length); - } - log_handler_ = &log_handler; disk_buffer_pool_ = &buffer_pool; @@ -863,14 +836,9 @@ RC BplusTreeHandler::create(LogHandler &log_handler, return RC::INTERNAL; } - char *pdata = header_frame->data(); - IndexFileHeader *file_header = (IndexFileHeader *)pdata; - file_header->attr_length = attr_length; - file_header->key_length = attr_length + sizeof(RID); - file_header->attr_type = attr_type; - file_header->internal_max_size = internal_max_size; - file_header->leaf_max_size = leaf_max_size; - file_header->root_page = BP_INVALID_PAGE_NUM; + char *pdata = header_frame->data(); + IndexFileHeader *file_header = (IndexFileHeader *)pdata; + file_header->init(index); // 取消记录日志的原因请参考下面的sync调用的地方。 // mtr.logger().init_header_page(header_frame, *file_header); @@ -887,8 +855,9 @@ RC BplusTreeHandler::create(LogHandler &log_handler, return RC::NOMEM; } - key_comparator_.init(file_header->attr_type, file_header->attr_length); - key_printer_.init(file_header->attr_type, file_header->attr_length); + index_meta_ = index; + key_comparator_.init(index); + key_printer_.init(index); /* 虽然我们针对B+树记录了WAL，但是我们记录的都是逻辑日志，并没有记录某个页面如何修改的物理日志。 @@ -912,7 +881,7 @@ RC BplusTreeHandler::open(LogHandler &log_handler, BufferPoolManager &bpm, const return RC::RECORD_OPENNED; } - DiskBufferPool *disk_buffer_pool = nullptr; + DiskBufferPool *disk_buffer_pool = nullptr; RC rc = bpm.open_file(log_handler, file_name, disk_buffer_pool); if (OB_FAIL(rc)) { @@ -959,8 +928,8 @@ RC BplusTreeHandler::open(LogHandler &log_handler, DiskBufferPool &buffer_pool) // close old page_handle buffer_pool.unpin_page(frame); - key_comparator_.init(file_header_.attr_type, file_header_.attr_length); - key_printer_.init(file_header_.attr_type, file_header_.attr_length); + key_comparator_.init(index_meta_); + key_printer_.init(index_meta_); LOG_INFO("Successfully open index"); return RC::SUCCESS; } @@ -978,7 +947,7 @@ RC BplusTreeHandler::close() RC BplusTreeHandler::print_leaf(Frame *frame) { BplusTreeMiniTransaction mtr(*this); - LeafIndexNodeHandler leaf_node(mtr, file_header_, frame); + LeafIndexNodeHandler leaf_node(mtr, file_header_, frame); LOG_INFO("leaf node: %s", to_string(leaf_node, key_printer_).c_str()); disk_buffer_pool_->unpin_page(frame); return RC::SUCCESS; @@ -986,7 +955,7 @@ RC BplusTreeHandler::print_leaf(Frame *frame) RC BplusTreeHandler::print_internal_node_recursive(Frame *frame) { - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; BplusTreeMiniTransaction mtr(*this); LOG_INFO("bplus tree. file header: %s", file_header_.to_string().c_str()); @@ -1060,8 +1029,8 @@ RC BplusTreeHandler::print_leafs() } BplusTreeMiniTransaction mtr(*this); - LatchMemo latch_memo = mtr.latch_memo(); - Frame *frame = nullptr; + LatchMemo latch_memo = mtr.latch_memo(); + Frame *frame = nullptr; RC rc = left_most_page(mtr, frame); if (OB_FAIL(rc)) { @@ -1100,9 +1069,9 @@ bool BplusTreeHandler::validate_node_recursive(BplusTreeMiniTransaction &mtr, Fr InternalIndexNodeHandler internal_node(mtr, file_header_, frame); result = internal_node.validate(key_comparator_, disk_buffer_pool_); for (int i = 0; result && i < internal_node.size(); i++) { - PageNum page_num = internal_node.value_at(i); + PageNum page_num = internal_node.value_at(i); Frame *child_frame = nullptr; - RC rc = mtr.latch_memo().get_page(page_num, child_frame); + RC rc = mtr.latch_memo().get_page(page_num, child_frame); if (OB_FAIL(rc)) { LOG_WARN("failed to fetch child page.page id=%d, rc=%d:%s", page_num, rc, strrc(rc)); result = false; @@ -1165,8 +1134,8 @@ bool BplusTreeHandler::validate_tree() } BplusTreeMiniTransaction mtr(*this); - LatchMemo &latch_memo = mtr.latch_memo(); - Frame *frame = nullptr; + LatchMemo &latch_memo = mtr.latch_memo(); + Frame *frame = nullptr; RC rc = latch_memo.get_page(file_header_.root_page, frame); // 这里仅仅调试使用，不加root锁 if (OB_FAIL(rc)) { @@ -1242,8 +1211,8 @@ RC BplusTreeHandler::crabing_protocal_fetch_page( BplusTreeMiniTransaction &mtr, BplusTreeOperationType op, PageNum page_num, bool is_root_node, Frame *&frame) { LatchMemo &latch_memo = mtr.latch_memo(); - bool readonly = (op == BplusTreeOperationType::READ); - const int memo_point = latch_memo.memo_point(); + bool readonly = (op == BplusTreeOperationType::READ); + const int memo_point = latch_memo.memo_point(); RC rc = latch_memo.get_page(page_num, frame); if (OB_FAIL(rc)) { @@ -1260,7 +1229,8 @@ RC BplusTreeHandler::crabing_protocal_fetch_page( return rc; } -RC BplusTreeHandler::insert_entry_into_leaf_node(BplusTreeMiniTransaction &mtr, Frame *frame, const char *key, const RID *rid) +RC BplusTreeHandler::insert_entry_into_leaf_node( + BplusTreeMiniTransaction &mtr, Frame *frame, const char *key, const RID *rid) { LeafIndexNodeHandler leaf_node(mtr, file_header_, frame); bool exists = false; // 该数据是否已经存在指定的叶子节点中了 @@ -1298,7 +1268,8 @@ RC BplusTreeHandler::insert_entry_into_leaf_node(BplusTreeMiniTransaction &mtr, return insert_entry_into_parent(mtr, frame, new_frame, new_index_node.key_at(0)); } -RC BplusTreeHandler::insert_entry_into_parent(BplusTreeMiniTransaction &mtr, Frame *frame, Frame *new_frame, const char *key) +RC BplusTreeHandler::insert_entry_into_parent( + BplusTreeMiniTransaction &mtr, Frame *frame, Frame *new_frame, const char *key) { RC rc = RC::SUCCESS; @@ -1310,7 +1281,7 @@ RC BplusTreeHandler::insert_entry_into_parent(BplusTreeMiniTransaction &mtr, Fra // create new root page Frame *root_frame = nullptr; - rc = disk_buffer_pool_->allocate_page(&root_frame); + rc = disk_buffer_pool_->allocate_page(&root_frame); if (OB_FAIL(rc)) { LOG_WARN("failed to allocate new root page. rc=%d:%s", rc, strrc(rc)); return rc; @@ -1430,16 +1401,17 @@ RC BplusTreeHandler::recover_update_root_page(BplusTreeMiniTransaction &mtr, Pag return RC::SUCCESS; } -RC BplusTreeHandler::recover_init_header_page(BplusTreeMiniTransaction &mtr, Frame *frame, const IndexFileHeader &header) +RC BplusTreeHandler::recover_init_header_page( + BplusTreeMiniTransaction &mtr, Frame *frame, const IndexFileHeader &header) { IndexFileHeader *file_header = reinterpret_cast(frame->data()); memcpy(file_header, &header, sizeof(IndexFileHeader)); - file_header_ = header; + file_header_ = header; header_dirty_ = false; frame->mark_dirty(); - key_comparator_.init(file_header_.attr_type, file_header_.attr_length); - key_printer_.init(file_header_.attr_type, file_header_.attr_length); + key_comparator_.init(index_meta_); + key_printer_.init(index_meta_); return RC::SUCCESS; } @@ -1732,7 +1704,8 @@ RC BplusTreeHandler::coalesce( } template -RC BplusTreeHandler::redistribute(BplusTreeMiniTransaction &mtr, Frame *neighbor_frame, Frame *frame, Frame *parent_frame, int index) +RC BplusTreeHandler::redistribute( + BplusTreeMiniTransaction &mtr, Frame *neighbor_frame, Frame *frame, Frame *parent_frame, int index) { InternalIndexNodeHandler parent_node(mtr, file_header_, parent_frame); IndexNodeHandlerType neighbor_node(mtr, file_header_, neighbor_frame); @@ -1821,9 +1794,7 @@ RC BplusTreeHandler::delete_entry(const char *user_key, const RID *rid) //////////////////////////////////////////////////////////////////////////////// -BplusTreeScanner::BplusTreeScanner(BplusTreeHandler &tree_handler) - : tree_handler_(tree_handler), mtr_(tree_handler) -{} +BplusTreeScanner::BplusTreeScanner(BplusTreeHandler &tree_handler) : tree_handler_(tree_handler), mtr_(tree_handler) {} BplusTreeScanner::~BplusTreeScanner() { close(); } @@ -1843,8 +1814,7 @@ RC BplusTreeScanner::open(const char *left_user_key, int left_len, bool left_inc // 校验输入的键值是否是合法范围 if (left_user_key && right_user_key) { - const auto &attr_comparator = tree_handler_.key_comparator_.attr_comparator(); - const int result = attr_comparator(left_user_key, right_user_key); + const int result = tree_handler_.key_comparator_.compare_key(left_user_key, right_user_key); if (result > 0 || // left < right // left == right but is (left,right)/[left,right) or (left,right] (result == 0 && (left_inclusive == false || right_inclusive == false))) { @@ -1859,7 +1829,7 @@ RC BplusTreeScanner::open(const char *left_user_key, int left_len, bool left_inc current_frame_ = nullptr; return RC::SUCCESS; } - + LOG_WARN("failed to find left most page. rc=%s", strrc(rc)); return rc; } @@ -1868,18 +1838,6 @@ RC BplusTreeScanner::open(const char *left_user_key, int left_len, bool left_inc } else { char *fixed_left_key = const_cast(left_user_key); - if (tree_handler_.file_header_.attr_type == AttrType::CHARS) { - bool should_inclusive_after_fix = false; - rc = fix_user_key(left_user_key, left_len, true /*greater*/, &fixed_left_key, &should_inclusive_after_fix); - if (OB_FAIL(rc)) { - LOG_WARN("failed to fix left user key. rc=%s", strrc(rc)); - return rc; - } - - if (should_inclusive_after_fix) { - left_inclusive = true; - } - } MemPoolItem::item_unique_ptr left_pkey; if (left_inclusive) { @@ -1933,19 +1891,7 @@ RC BplusTreeScanner::open(const char *left_user_key, int left_len, bool left_inc right_key_ = nullptr; } else { - char *fixed_right_key = const_cast(right_user_key); - bool should_include_after_fix = false; - if (tree_handler_.file_header_.attr_type == AttrType::CHARS) { - rc = fix_user_key(right_user_key, right_len, false /*want_greater*/, &fixed_right_key, &should_include_after_fix); - if (OB_FAIL(rc)) { - LOG_WARN("failed to fix right user key. rc=%s", strrc(rc)); - return rc; - } - - if (should_include_after_fix) { - right_inclusive = true; - } - } + char *fixed_right_key = const_cast(right_user_key); if (right_inclusive) { right_key_ = tree_handler_.make_key(fixed_right_key, *RID::max()); } else { @@ -2045,53 +1991,14 @@ RC BplusTreeScanner::close() return RC::SUCCESS; } -RC BplusTreeScanner::fix_user_key( - const char *user_key, int key_len, bool want_greater, char **fixed_key, bool *should_inclusive) +RC IndexFileHeader::init(const IndexMeta &index) { - if (nullptr == fixed_key || nullptr == should_inclusive) { - return RC::INVALID_ARGUMENT; - } - - // 这里很粗暴，变长字段才需要做调整，其它默认都不需要做调整 - assert(tree_handler_.file_header_.attr_type == AttrType::CHARS); - assert(strlen(user_key) >= static_cast(key_len)); + root_page = BP_INVALID_PAGE_NUM; + attr_length = index.fields_total_len(); + key_length = attr_length + static_cast(sizeof(RID)); - *should_inclusive = false; - - int32_t attr_length = tree_handler_.file_header_.attr_length; - char *key_buf = new char[attr_length]; - if (nullptr == key_buf) { - return RC::NOMEM; - } - - if (key_len <= attr_length) { - memcpy(key_buf, user_key, key_len); - memset(key_buf + key_len, 0, attr_length - key_len); - - *fixed_key = key_buf; - return RC::SUCCESS; - } - - // key_len > attr_length - memcpy(key_buf, user_key, attr_length); - - char c = user_key[attr_length]; - if (c == 0) { - *fixed_key = key_buf; - return RC::SUCCESS; - } - - // 扫描 >=/> user_key 的数据 - // 示例：>=/> ABCD1 的数据，attr_length=4, - // 等价于扫描 >= ABCE 的数据 - // 如果是扫描 <=/< user_key的数据 - // 示例：<=/< ABCD1 <==> <= ABCD (attr_length=4) - // NOTE: 假设都是普通的ASCII字符，不包含二进制字符，使用char不会溢出 - *should_inclusive = true; - if (want_greater) { - key_buf[attr_length - 1]++; - } + internal_max_size = calc_internal_page_capacity(index.fields_total_len()); + leaf_max_size = calc_leaf_page_capacity(index.fields_total_len()); - *fixed_key = key_buf; return RC::SUCCESS; } diff --git a/src/observer/storage/index/bplus_tree.h b/src/observer/storage/index/bplus_tree.h index dd752056..515ac79d 100644 --- a/src/observer/storage/index/bplus_tree.h +++ b/src/observer/storage/index/bplus_tree.h @@ -56,6 +56,10 @@ enum class BplusTreeOperationType class AttrComparator { public: + AttrComparator() = default; + + AttrComparator(AttrType type, int length) { init(type, length); } + void init(AttrType type, int length) { attr_type_ = type; @@ -89,50 +93,54 @@ class AttrComparator class KeyComparator { public: - void init(AttrType type, int length) { attr_comparator_.init(type, length); } + void init(const IndexMeta &index) + { + index_ = index; + for (const auto &i : index.fields()) { + attr_comparator_.emplace_back(i.type(), (i.len())); + } + } - const AttrComparator &attr_comparator() const { return attr_comparator_; } + int compare_key(const char *v1, const char *v2) const + { + auto field_number = index_.fields().size(); + auto &fields_offset = index_.fields_offset(); + for (size_t i = 0; i < field_number; i++) { + int offset = fields_offset[i]; + auto &field = index_.fields()[i]; + if (field.nullable()) { + bool v1_is_null = v1[offset + field.len() - 1] == '1'; + bool v2_is_null = v2[offset + field.len() - 1] == '1'; + if (v1_is_null) { + return -1; + } + if (v2_is_null) { + return 1; + } + } + int result = attr_comparator_[i](v1 + offset, v2 + offset); + if (result != 0) { + return result; + } + } + return 0; + } int operator()(const char *v1, const char *v2) const { - int result = attr_comparator_(v1, v2); + auto result = compare_key(v1, v2); if (result != 0) { return result; } - const RID *rid1 = (const RID *)(v1 + attr_comparator_.attr_length()); - const RID *rid2 = (const RID *)(v2 + attr_comparator_.attr_length()); + const RID *rid1 = (const RID *)(v1 + index_.fields_total_len()); + const RID *rid2 = (const RID *)(v2 + index_.fields_total_len()); return RID::compare(rid1, rid2); } private: - AttrComparator attr_comparator_; -}; - -/** - * @brief 属性打印,调试使用(BplusTree) - * @ingroup BPlusTree - */ -class AttrPrinter -{ -public: - void init(AttrType type, int length) - { - attr_type_ = type; - attr_length_ = length; - } - - int attr_length() const { return attr_length_; } - - string operator()(const char *v) const - { - Value value(attr_type_, const_cast(v), attr_length_); - return value.to_string(); - } - -private: - AttrType attr_type_; - int attr_length_; + IndexMeta index_; + vector attr_comparator_; }; /** @@ -142,22 +150,20 @@ class AttrPrinter class KeyPrinter { public: - void init(AttrType type, int length) { attr_printer_.init(type, length); } - - const AttrPrinter &attr_printer() const { return attr_printer_; } + void init(const IndexMeta &index) { index_ = index; } string operator()(const char *v) const { stringstream ss; - ss << "{key:" << attr_printer_(v) << ","; + ss << "{key:" << index_.to_string() << ","; - const RID *rid = (const RID *)(v + attr_printer_.attr_length()); + const RID *rid = (const RID *)(v + index_.fields_total_len()); ss << "rid:{" << rid->to_string() << "}}"; return ss.str(); } private: - AttrPrinter attr_printer_; + IndexMeta index_; }; /** @@ -168,17 +174,13 @@ class KeyPrinter */ struct IndexFileHeader { - IndexFileHeader() - { - memset(this, 0, sizeof(IndexFileHeader)); - root_page = BP_INVALID_PAGE_NUM; - } - PageNum root_page; ///< 根节点在磁盘中的页号 - int32_t internal_max_size; ///< 内部节点最大的键值对数 - int32_t leaf_max_size; ///< 叶子节点最大的键值对数 - int32_t attr_length; ///< 键值的长度 - int32_t key_length; ///< attr length + sizeof(RID) - AttrType attr_type; ///< 键值的类型 + RC init(const IndexMeta &index); + + PageNum root_page = BP_INVALID_PAGE_NUM; ///< 根节点在磁盘中的页号 + int32_t internal_max_size; ///< 内部节点最大的键值对数 + int32_t leaf_max_size; ///< 叶子节点最大的键值对数 + int32_t attr_length; ///< 字段部分的长度 + int32_t key_length; ///< 键的长度 const string to_string() const { @@ -186,7 +188,6 @@ struct IndexFileHeader ss << "attr_length:" << attr_length << "," << "key_length:" << key_length << "," - << "attr_type:" << attr_type_to_string(attr_type) << "," << "root_page:" << root_page << "," << "internal_max_size:" << internal_max_size << "," << "leaf_max_size:" << leaf_max_size << ";"; @@ -459,10 +460,8 @@ class BplusTreeHandler * @param internal_max_size 内部节点最大大小 * @param leaf_max_size 叶子节点最大大小 */ - RC create(LogHandler &log_handler, BufferPoolManager &bpm, const char *file_name, AttrType attr_type, int attr_length, - int internal_max_size = -1, int leaf_max_size = -1); - RC create(LogHandler &log_handler, DiskBufferPool &buffer_pool, AttrType attr_type, int attr_length, - int internal_max_size = -1, int leaf_max_size = -1); + RC create(LogHandler &log_handler, BufferPoolManager &bpm, const char *file_name, const IndexMeta &index); + RC create(LogHandler &log_handler, DiskBufferPool &buffer_pool, const IndexMeta &index); /** * @brief 打开一个B+树 @@ -640,6 +639,7 @@ class BplusTreeHandler DiskBufferPool *disk_buffer_pool_ = nullptr; /// 磁盘缓冲池 bool header_dirty_ = false; /// 是否需要更新头页面 IndexFileHeader file_header_; + IndexMeta index_meta_; // 在调整根节点时，需要加上这个锁。 // 这个锁可以使用递归读写锁，但是这里偷懒先不改 @@ -696,11 +696,6 @@ class BplusTreeScanner RC close(); private: - /** - * 如果key的类型是CHARS, 扩展或缩减user_key的大小刚好是schema中定义的大小 - */ - RC fix_user_key(const char *user_key, int key_len, bool want_greater, char **fixed_key, bool *should_inclusive); - void fetch_item(RID &rid); /** diff --git a/src/observer/storage/index/bplus_tree_index.cpp b/src/observer/storage/index/bplus_tree_index.cpp index 8794cfbc..10eadcb4 100644 --- a/src/observer/storage/index/bplus_tree_index.cpp +++ b/src/observer/storage/index/bplus_tree_index.cpp @@ -22,57 +22,57 @@ BplusTreeIndex::~BplusTreeIndex() noexcept { close(); } RC BplusTreeIndex::create(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) { if (inited_) { - LOG_WARN("Failed to create index due to the index has been created before. file_name:%s, index:%s, field:%s", - file_name, index_meta.name(), index_meta.field()); + LOG_WARN("Failed to create index due to the index has been created before. file_name:%s, index:%s", + file_name, index_meta.to_string().c_str()); return RC::RECORD_OPENNED; } - Index::init(index_meta, field_meta); + Index::init(index_meta); BufferPoolManager &bpm = table->db()->buffer_pool_manager(); - RC rc = index_handler_.create(table->db()->log_handler(), bpm, file_name, field_meta.type(), field_meta.len()); + RC rc = index_handler_.create(table->db()->log_handler(), bpm, file_name, index_meta); if (RC::SUCCESS != rc) { - LOG_WARN("Failed to create index_handler, file_name:%s, index:%s, field:%s, rc:%s", - file_name, index_meta.name(), index_meta.field(), strrc(rc)); + LOG_WARN("Failed to create index_handler, file_name:%s, index:%s, rc:%s", + file_name, index_meta.to_string().c_str(), strrc(rc)); return rc; } inited_ = true; table_ = table; - LOG_INFO("Successfully create index, file_name:%s, index:%s, field:%s", - file_name, index_meta.name(), index_meta.field()); + LOG_INFO("Successfully create index, file_name:%s, index:%s", + file_name, index_meta.to_string().c_str()); return RC::SUCCESS; } RC BplusTreeIndex::open(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) { if (inited_) { - LOG_WARN("Failed to open index due to the index has been initedd before. file_name:%s, index:%s, field:%s", - file_name, index_meta.name(), index_meta.field()); + LOG_WARN("Failed to open index due to the index has been initedd before. file_name:%s, index:%s", + file_name, index_meta.to_string().c_str()); return RC::RECORD_OPENNED; } - Index::init(index_meta, field_meta); + Index::init(index_meta); BufferPoolManager &bpm = table->db()->buffer_pool_manager(); - RC rc = index_handler_.open(table->db()->log_handler(), bpm, file_name); + RC rc = index_handler_.open(table->db()->log_handler(), bpm, file_name); if (RC::SUCCESS != rc) { - LOG_WARN("Failed to open index_handler, file_name:%s, index:%s, field:%s, rc:%s", - file_name, index_meta.name(), index_meta.field(), strrc(rc)); + LOG_WARN("Failed to open index_handler, file_name:%s, index:%s, rc:%s", + file_name, index_meta.to_string().c_str(), strrc(rc)); return rc; } inited_ = true; table_ = table; - LOG_INFO("Successfully open index, file_name:%s, index:%s, field:%s", - file_name, index_meta.name(), index_meta.field()); + LOG_INFO("Successfully open index, file_name:%s, index:%s", + file_name, index_meta.to_string().c_str()); return RC::SUCCESS; } RC BplusTreeIndex::close() { if (inited_) { - LOG_INFO("Begin to close index, index:%s, field:%s", index_meta_.name(), index_meta_.field()); + LOG_INFO("Begin to close index, index:%s", index_meta_.to_string().c_str()); index_handler_.close(); inited_ = false; } @@ -82,12 +82,25 @@ RC BplusTreeIndex::close() RC BplusTreeIndex::insert_entry(const char *record, const RID *rid) { - return index_handler_.insert_entry(record + field_meta_.offset(), rid); + char *entry = index_meta_.make_entry_from_record(record); + if (index_meta_.unique()) { + list entries; + RC rc = index_handler_.get_entry(entry, index_meta_.fields_total_len(), entries); + if (OB_FAIL(rc)) { + return rc; + } + if (!entries.empty()) { + return RC::RECORD_DUPLICATE_KEY; + } + } + + return index_handler_.insert_entry(entry, rid); } RC BplusTreeIndex::delete_entry(const char *record, const RID *rid) { - return index_handler_.delete_entry(record + field_meta_.offset(), rid); + char *entry = index_meta_.make_entry_from_record(record); + return index_handler_.delete_entry(entry, rid); } IndexScanner *BplusTreeIndex::create_scanner( diff --git a/src/observer/storage/index/bplus_tree_index.h b/src/observer/storage/index/bplus_tree_index.h index db0bbc99..773ea7d7 100644 --- a/src/observer/storage/index/bplus_tree_index.h +++ b/src/observer/storage/index/bplus_tree_index.h @@ -27,8 +27,10 @@ class BplusTreeIndex : public Index BplusTreeIndex() = default; virtual ~BplusTreeIndex() noexcept; - RC create(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) override; - RC open(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) override; + RC create( + Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta = {}) override; + RC open(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta = {}) override; + RC close(); RC insert_entry(const char *record, const RID *rid) override; diff --git a/src/observer/storage/index/bplus_tree_log.cpp b/src/observer/storage/index/bplus_tree_log.cpp index e2ee32ff..573d7b9b 100644 --- a/src/observer/storage/index/bplus_tree_log.cpp +++ b/src/observer/storage/index/bplus_tree_log.cpp @@ -182,13 +182,14 @@ RC BplusTreeLogger::redo(BufferPoolManager &bpm, const LogEntry &entry) return rc; } -RC BplusTreeLogger::__redo(LSN lsn, BplusTreeMiniTransaction &mtr, BplusTreeHandler &tree_handler, Deserializer &redo_buffer) +RC BplusTreeLogger::__redo( + LSN lsn, BplusTreeMiniTransaction &mtr, BplusTreeHandler &tree_handler, Deserializer &redo_buffer) { need_log_ = false; DEFER(need_log_ = true); - RC rc = RC::SUCCESS; + RC rc = RC::SUCCESS; vector frames; while (redo_buffer.remain() > 0) { unique_ptr entry; @@ -202,8 +203,8 @@ RC BplusTreeLogger::__redo(LSN lsn, BplusTreeMiniTransaction &mtr, BplusTreeHand if (frame != nullptr) { if (frame->lsn() >= lsn) { LOG_TRACE("no need to redo. frame=%p:%s, redo lsn=%ld", frame, frame->to_string().c_str(), lsn); - frame->unpin(); - continue; + frame->unpin(); + continue; } else { frames.push_back(frame); } @@ -273,7 +274,7 @@ BplusTreeMiniTransaction::~BplusTreeMiniTransaction() if (nullptr == operation_result_) { return; } - + if (OB_SUCC(*operation_result_)) { commit(); } else { diff --git a/src/observer/storage/index/bplus_tree_log_entry.cpp b/src/observer/storage/index/bplus_tree_log_entry.cpp index 615bb52b..1f41c8e0 100644 --- a/src/observer/storage/index/bplus_tree_log_entry.cpp +++ b/src/observer/storage/index/bplus_tree_log_entry.cpp @@ -363,7 +363,7 @@ LeafInitEmptyLogEntryHandler::LeafInitEmptyLogEntryHandler(Frame *frame) RC LeafInitEmptyLogEntryHandler::redo(BplusTreeMiniTransaction &mtr, BplusTreeHandler &tree_handler) { LeafIndexNodeHandler leaf_handler(mtr, tree_handler.file_header(), frame()); - RC rc = leaf_handler.init_empty(); + RC rc = leaf_handler.init_empty(); return rc; } @@ -494,7 +494,7 @@ RC InternalCreateNewRootLogEntryHandler::deserialize( RC InternalCreateNewRootLogEntryHandler::redo(BplusTreeMiniTransaction &mtr, BplusTreeHandler &tree_handler) { InternalIndexNodeHandler internal_handler(mtr, tree_handler.file_header(), frame()); - RC rc = internal_handler.create_new_root(first_page_num_, key_.data(), page_num_); + RC rc = internal_handler.create_new_root(first_page_num_, key_.data(), page_num_); return rc; } diff --git a/src/observer/storage/index/index.cpp b/src/observer/storage/index/index.cpp index 6fa51531..b364f7f9 100644 --- a/src/observer/storage/index/index.cpp +++ b/src/observer/storage/index/index.cpp @@ -14,9 +14,8 @@ See the Mulan PSL v2 for more details. */ #include "storage/index/index.h" -RC Index::init(const IndexMeta &index_meta, const FieldMeta &field_meta) +RC Index::init(const IndexMeta &index_meta) { index_meta_ = index_meta; - field_meta_ = field_meta; return RC::SUCCESS; } diff --git a/src/observer/storage/index/index.h b/src/observer/storage/index/index.h index 790c34cd..8f1a18e4 100644 --- a/src/observer/storage/index/index.h +++ b/src/observer/storage/index/index.h @@ -44,12 +44,13 @@ class Index { return RC::UNSUPPORTED; } + virtual RC open(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) { return RC::UNSUPPORTED; } - virtual bool is_vector_index() { return false; } + virtual bool is_vector_index() { return index_meta_.index_type() == IndexType::VectorIVFFlatIndex; } const IndexMeta &index_meta() const { return index_meta_; } @@ -80,7 +81,10 @@ class Index * @param right_inclusive 是否包含右边界 */ virtual IndexScanner *create_scanner(const char *left_key, int left_len, bool left_inclusive, const char *right_key, - int right_len, bool right_inclusive) = 0; + int right_len, bool right_inclusive) + { + return nullptr; + } /** * @brief 同步索引数据到磁盘 @@ -89,11 +93,10 @@ class Index virtual RC sync() = 0; protected: - RC init(const IndexMeta &index_meta, const FieldMeta &field_meta); + RC init(const IndexMeta &index_meta); protected: IndexMeta index_meta_; ///< 索引的元数据 - FieldMeta field_meta_; ///< 当前实现仅考虑一个字段的索引 }; /** diff --git a/src/observer/storage/index/index_meta.cpp b/src/observer/storage/index/index_meta.cpp index 32604b9f..cd004aae 100644 --- a/src/observer/storage/index/index_meta.cpp +++ b/src/observer/storage/index/index_meta.cpp @@ -8,64 +8,86 @@ EITHER EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO NON-INFRINGEMENT, MERCHANTABILITY OR FIT FOR A PARTICULAR PURPOSE. See the Mulan PSL v2 for more details. */ -// -// Created by Wangyunlai.wyl on 2021/5/18. -// +#include "index_meta.h" -#include "storage/index/index_meta.h" -#include "common/lang/string.h" -#include "common/log/log.h" -#include "storage/field/field_meta.h" -#include "storage/table/table_meta.h" -#include "json/json.h" - -const static Json::StaticString FIELD_NAME("name"); -const static Json::StaticString FIELD_FIELD_NAME("field_name"); - -RC IndexMeta::init(const char *name, const FieldMeta &field) +RC IndexMeta::init(const char *name, IndexType index_type, const vector &fields, bool unique) { - if (common::is_blank(name)) { - LOG_ERROR("Failed to init index, name is empty."); - return RC::INVALID_ARGUMENT; + name_ = name; + index_type_ = index_type; + fields_total_len_ = 0; + fields_ = fields; + unique_ = unique; + + for (auto &field : fields) { + fields_offset_.emplace_back(fields_total_len_); + fields_total_len_ += field.len(); } - name_ = name; - field_ = field.name(); return RC::SUCCESS; } -void IndexMeta::to_json(Json::Value &json_value) const +string IndexMeta::to_string() const { - json_value[FIELD_NAME] = name_; - json_value[FIELD_FIELD_NAME] = field_; + std::ostringstream oss; + oss << "Index Name: " << name_ << ", Fields: ["; + for (size_t i = 0; i < fields_.size(); ++i) { + oss << fields_[i].name(); + if (i < fields_.size() - 1) { + oss << ", "; + } + } + oss << "], Total Length: " << fields_total_len_; + return oss.str(); } -RC IndexMeta::from_json(const TableMeta &table, const Json::Value &json_value, IndexMeta &index) +void IndexMeta::to_json(Json::Value &json_value) const { - const Json::Value &name_value = json_value[FIELD_NAME]; - const Json::Value &field_value = json_value[FIELD_FIELD_NAME]; - if (!name_value.isString()) { - LOG_ERROR("Index name is not a string. json value=%s", name_value.toStyledString().c_str()); - return RC::INTERNAL; + json_value["name"] = name_; + json_value["index_type"] = static_cast(index_type_); + json_value["fields_total_len"] = fields_total_len_; + json_value["unique"] = unique_; + + Json::Value fields_json(Json::arrayValue); + for (const auto &field : fields_) { + Json::Value field_json; + field.to_json(field_json); + fields_json.append(field_json); } + json_value["fields"] = fields_json; - if (!field_value.isString()) { - LOG_ERROR("Field name of index [%s] is not a string. json value=%s", - name_value.asCString(), field_value.toStyledString().c_str()); - return RC::INTERNAL; + Json::Value offsets_json(Json::arrayValue); + for (const auto &offset : fields_offset_) { + offsets_json.append(offset); } + json_value["fields_offset"] = offsets_json; +} - const FieldMeta *field = table.field(field_value.asCString()); - if (nullptr == field) { - LOG_ERROR("Deserialize index [%s]: no such field: %s", name_value.asCString(), field_value.asCString()); - return RC::SCHEMA_FIELD_MISSING; +RC IndexMeta::from_json(const Json::Value &json_value, IndexMeta &index) +{ + if (!json_value.isMember("name") || !json_value.isMember("index_type") || !json_value.isMember("fields") || + !json_value.isMember("fields_total_len") || !json_value.isMember("fields_offset")) { + return RC::INVALID_ARGUMENT; } - return index.init(name_value.asCString(), *field); -} + index.name_ = json_value["name"].asString(); + index.index_type_ = static_cast(json_value["index_type"].asInt()); + index.fields_total_len_ = json_value["fields_total_len"].asInt(); + index.unique_ = json_value["unique"].asBool(); -const char *IndexMeta::name() const { return name_.c_str(); } + const Json::Value &fields_json = json_value["fields"]; + for (const auto &field_json : fields_json) { + FieldMeta field; + RC rc = FieldMeta::from_json(field_json, field); + if (rc != RC::SUCCESS) { + return rc; + } + index.fields_.push_back(field); + } -const char *IndexMeta::field() const { return field_.c_str(); } + const Json::Value &offsets_json = json_value["fields_offset"]; + for (const auto &offset : offsets_json) { + index.fields_offset_.push_back(offset.asInt()); + } -void IndexMeta::desc(ostream &os) const { os << "index name=" << name_ << ", field=" << field_; } \ No newline at end of file + return RC::SUCCESS; +} diff --git a/src/observer/storage/index/index_meta.h b/src/observer/storage/index/index_meta.h index 04d10642..6f420252 100644 --- a/src/observer/storage/index/index_meta.h +++ b/src/observer/storage/index/index_meta.h @@ -16,6 +16,9 @@ See the Mulan PSL v2 for more details. */ #include "common/rc.h" #include "common/lang/string.h" +#include "storage/field/field_meta.h" + +#include class TableMeta; class FieldMeta; @@ -35,19 +38,43 @@ class IndexMeta public: IndexMeta() = default; - RC init(const char *name, const FieldMeta &field); + [[nodiscard]] RC init(const char *name, IndexType index_type, const vector &fields, bool unique = false); -public: - const char *name() const; - const char *field() const; + void desc(ostream &os) const { os << to_string(); } - void desc(ostream &os) const; + [[nodiscard]] string to_string() const; -public: - void to_json(Json::Value &json_value) const; - static RC from_json(const TableMeta &table, const Json::Value &json_value, IndexMeta &index); + void to_json(Json::Value &json_value) const; + + static RC from_json(const Json::Value &json_value, IndexMeta &index); + + [[nodiscard]] char *make_entry_from_record(const char *record) + { + char *entry = new char[fields_total_len_]; + make_entry_from_record(entry, record); + return entry; + } + + void make_entry_from_record(char *entry, const char *record) + { + for (size_t i = 0; i < fields_.size(); i++) { + auto &field = fields_[i]; + memcpy(entry + fields_offset_[i], record + field.offset(), field.len()); + } + } + + [[nodiscard]] const char *name() const { return name_.c_str(); } + IndexType index_type() const { return index_type_; } + [[nodiscard]] int fields_total_len() const { return fields_total_len_; } + [[nodiscard]] const vector &fields() const { return fields_; } + [[nodiscard]] bool unique() const { return unique_; } + [[nodiscard]] const vector &fields_offset() const { return fields_offset_; } -protected: - string name_; // index's name - string field_; // field's name +private: + string name_; + IndexType index_type_; + int fields_total_len_ = 0; + vector fields_offset_; + vector fields_; + bool unique_; }; diff --git a/src/observer/storage/index/ivfflat_index.cpp b/src/observer/storage/index/ivfflat_index.cpp new file mode 100644 index 00000000..1857af12 --- /dev/null +++ b/src/observer/storage/index/ivfflat_index.cpp @@ -0,0 +1,234 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/18 * + * @Description : ivfflat index source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include + +#include "storage/index/ivfflat_index.h" + +#define MAX_ITERATIONS 5 + +using Vector = std::vector; + +void vector_add(Vector &a, const Vector &b) +{ + for (size_t i = 0; i < a.size(); i++) { + a[i] += b[i]; + } +} + +void vector_scalar_div(Vector &a, float x) +{ + for (auto &y : a) { + y /= x; + } +} + +// 计算两个向量之间的距离 +float compute_distance(const Vector &a, const Vector &b, NormalFunctionType dist_fn) +{ + switch (dist_fn) { + case NormalFunctionType::L2_DISTANCE: return builtin::l2_distance(a, b); + case NormalFunctionType::COSINE_DISTANCE: return builtin::cosine_distance(a, b); + case NormalFunctionType::INNER_PRODUCT: return builtin::inner_product(a, b); + default: return 0; + } +} + +// 找到离给定向量最近的质心 +auto find_centroid(const Vector &vec, const std::vector ¢roids, NormalFunctionType dist_fn) -> size_t +{ + size_t best_index = 0; + float min_distance = std::numeric_limits::max(); + for (size_t i = 0; i < centroids.size(); i++) { + float distance = compute_distance(vec, centroids[i], dist_fn); + if (distance < min_distance) { + min_distance = distance; + best_index = i; + } + } + return best_index; +} + +// 根据数据和旧质心计算新质心 +auto find_centroids(const std::vector> &data, const std::vector ¢roids, + NormalFunctionType dist_fn) -> std::vector +{ + std::vector new_centroids(centroids.size(), Vector(centroids[0].size(), 0.0)); + std::vector counts(centroids.size(), 0); + + // 聚类点累加 + for (const auto &[vec, rid] : data) { + size_t index = find_centroid(vec, centroids, dist_fn); + vector_add(new_centroids[index], vec); + counts[index]++; + } + + // 计算平均值作为新质心 + for (size_t i = 0; i < new_centroids.size(); i++) { + if (counts[i] > 0) { + vector_scalar_div(new_centroids[i], counts[i]); + } else { + new_centroids[i] = centroids[i]; // 无数据点的运行质心 + } + } + return new_centroids; +} + +RC IvfflatIndex::create(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) +{ + if (inited_) { + LOG_WARN("Failed to create index due to the index has been created before. file_name:%s, index:%s", + file_name, index_meta.to_string().c_str()); + return RC::RECORD_OPENNED; + } + + Index::init(index_meta); + + // 暂时先不支持持久化 + // BufferPoolManager &bpm = table->db()->buffer_pool_manager(); + // RC rc = index_handler_.create(table->db()->log_handler(), bpm, file_name, index_meta); + // if (RC::SUCCESS != rc) { + // LOG_WARN("Failed to create index_handler, file_name:%s, index:%s, rc:%s", + // file_name, index_meta.to_string().c_str(), strrc(rc)); + // return rc; + // } + + inited_ = true; + table_ = table; + field_meta_ = field_meta; + LOG_INFO("Successfully create index, file_name:%s, index:%s", + file_name, index_meta.to_string().c_str()); + return RC::SUCCESS; +} + +RC IvfflatIndex::open(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) +{ + return RC::SUCCESS; +} + +RC IvfflatIndex::close() { return RC::SUCCESS; } + +RC IvfflatIndex::build_index( + std::vector> &initial_data, NormalFunctionType distance_fn, const std::vector &options) +{ + // 初始化距离度量函数 + distance_fn_ = distance_fn; + + // 初始化参数 + lists_ = options[0]; + probes_ = options[1]; + + if (initial_data.empty()) { + return RC::SUCCESS; + } + + std::mt19937 gen(std::random_device{}()); + std::uniform_int_distribution<> dis(0, initial_data.size() - 1); + + // 随机选择初始质心 + centroids_.resize(lists_); + for (auto ¢roid : centroids_) { + centroid = initial_data[dis(gen)].first; + } + + // 迭代更新质心直到收敛 + size_t cnt = 0; // 迭代次数 + bool changed = true; + while (changed && cnt < MAX_ITERATIONS) { + changed = false; + auto new_centroids = find_centroids(initial_data, centroids_, distance_fn_); + + // 检查质心是否发生变化 + for (size_t i = 0; i < lists_; i++) { + if (compute_distance(new_centroids[i], centroids_[i], distance_fn_) > 0.01) { + changed = true; + } + centroids_[i] = std::move(new_centroids[i]); + } + + ++cnt; + } + + // 将数据点分配到最近的质心 + centroids_buckets_.resize(lists_); + for (auto &[vec, rid] : initial_data) { + size_t index = find_centroid(vec, centroids_, distance_fn_); + centroids_buckets_[index].emplace_back(std::move(vec), rid); + } + + return RC::SUCCESS; +} + +std::vector IvfflatIndex::ann_search(const std::vector &base_vector, size_t limit) +{ + std::vector> centroid_distances; + + // 计算与所有簇中心的距离 + for (size_t i = 0; i < centroids_.size(); i++) { + float dist = compute_distance(base_vector, centroids_[i], distance_fn_); + centroid_distances.emplace_back(dist, i); + } + + // 找到最近的 probes_ 个簇 + std::sort(centroid_distances.begin(), + centroid_distances.end(), + [&](const std::pair &a, const std::pair &b) { + return a.first < b.first; // 按距离升序排序 + }); + + std::vector> candidates; + + // 探测最近的 probes_ 个簇 + for (size_t i = 0; i < probes_; i++) { + size_t cluster_index = centroid_distances[i].second; + for (const auto &[vec, rid] : centroids_buckets_[cluster_index]) { + float dist = compute_distance(base_vector, vec, distance_fn_); + candidates.emplace_back(dist, rid); + } + } + + // 根据距离排序并选取前 limit 个结果 + std::sort(candidates.begin(), candidates.end(), [&](const std::pair &a, const std::pair &b) { + return a.first < b.first; // 按距离升序排序 + }); + + auto size = std::min(candidates.size(), limit); + std::vector result(size); + for (size_t i = 0; i < size; i++) { + result[i] = candidates[i].second; + } + + return result; +} + +Vector IvfflatIndex::get_vector(const char *record) +{ + int size = field_meta_.len() / sizeof(float); + Vector vector(size); + // 直接将记录位置调整为适当的浮点数指针 + const float *data = reinterpret_cast(record + field_meta_.offset()); + // 使用标准库函数来进行快速的内存复制 + std::copy(data, data + size, vector.begin()); + return vector; +} + +RC IvfflatIndex::insert_entry(const char *record, const RID *rid) +{ + Vector vector = get_vector(record); + size_t index = find_centroid(vector, centroids_, this->distance_fn_); + centroids_buckets_[index].emplace_back(std::move(vector), *rid); + return RC::SUCCESS; +} + +RC IvfflatIndex::delete_entry(const char *record, const RID *rid) { return RC::SUCCESS; } + +RC IvfflatIndex::sync() { return RC::SUCCESS; } diff --git a/src/observer/storage/index/ivfflat_index.h b/src/observer/storage/index/ivfflat_index.h index cb09f47a..644c4b42 100644 --- a/src/observer/storage/index/ivfflat_index.h +++ b/src/observer/storage/index/ivfflat_index.h @@ -11,6 +11,7 @@ See the Mulan PSL v2 for more details. */ #pragma once #include "storage/index/index.h" +#include "sql/builtin/builtin.h" /** * @brief ivfflat 向量索引 @@ -19,31 +20,46 @@ See the Mulan PSL v2 for more details. */ class IvfflatIndex : public Index { public: - IvfflatIndex(){}; - virtual ~IvfflatIndex() noexcept {}; + using Vector = std::vector; - RC create(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) - { - return RC::UNIMPLEMENTED; - }; - RC open(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) - { + IvfflatIndex() = default; + ~IvfflatIndex() noexcept override = default; - return RC::UNIMPLEMENTED; - }; + RC create(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) override; - vector ann_search(const vector &base_vector, size_t limit) { return vector(); } + RC open(Table *table, const char *file_name, const IndexMeta &index_meta, const FieldMeta &field_meta) override; - RC close() { return RC::UNIMPLEMENTED; } + RC close(); - RC insert_entry(const char *record, const RID *rid) override { return RC::UNIMPLEMENTED; }; - RC delete_entry(const char *record, const RID *rid) override { return RC::UNIMPLEMENTED; }; + std::vector ann_search(const std::vector &base_vector, size_t limit); - RC sync() override { return RC::UNIMPLEMENTED; }; + Vector get_vector(const char *record); + + RC insert_entry(const char *record, const RID *rid) override; + + RC delete_entry(const char *record, const RID *rid) override; + + RC sync() override; + + RC build_index(std::vector> &initial_data, NormalFunctionType distance_fn, + const std::vector &options); + + NormalFunctionType distance_fn() { return distance_fn_; } private: bool inited_ = false; Table *table_ = nullptr; int lists_ = 1; int probes_ = 1; + + FieldMeta field_meta_; // 建立索引字段元数据 + + // 距离度量函数 + NormalFunctionType distance_fn_; + + // 质心 + std::vector centroids_; + + // 每个质心周围的点 + std::vector>> centroids_buckets_; }; diff --git a/src/observer/storage/record/record.h b/src/observer/storage/record/record.h index 6749804d..64886b43 100644 --- a/src/observer/storage/record/record.h +++ b/src/observer/storage/record/record.h @@ -25,7 +25,10 @@ See the Mulan PSL v2 for more details. */ #include "storage/field/field_meta.h" #include "storage/index/index_meta.h" +#include + class Field; +class BaseTable; /** * @brief 标识一个记录的位置 @@ -101,20 +104,24 @@ class Record { public: Record() = default; - ~Record() + ~Record() { reset(); } + + void reset() { if (owner_ && data_ != nullptr) { free(data_); - data_ = nullptr; + owner_ = false; + data_ = nullptr; } } Record(const Record &other) { - rid_ = other.rid_; - data_ = other.data_; - len_ = other.len_; - owner_ = other.owner_; + rid_ = other.rid_; + base_rids_ = other.base_rids_; + data_ = other.data_; + len_ = other.len_; + owner_ = other.owner_; if (other.owner_) { char *tmp = (char *)malloc(other.len_); @@ -131,18 +138,32 @@ class Record } if (!owner_ || len_ != other.len_) { - this->~Record(); + reset(); new (this) Record(other); return *this; } - this->rid_ = other.rid_; + this->rid_ = other.rid_; + this->base_rids_ = other.base_rids_; memcpy(data_, other.data_, other.len_); return *this; } + Record clone() + { + Record new_record; + new_record.rid_ = this->rid_; + new_record.base_rids_ = this->base_rids_; + new_record.len_ = this->len_; + new_record.data_ = (char *)malloc(this->len_); + memcpy(new_record.data_, this->data_, this->len_); + new_record.owner_ = true; + return new_record; + } + Record(Record &&other) { - rid_ = other.rid_; + rid_ = other.rid_; + base_rids_ = std::move(other.base_rids_); if (!other.owner_) { data_ = other.data_; @@ -165,7 +186,7 @@ class Record return *this; } - this->~Record(); + reset(); new (this) Record(std::move(other)); return *this; } @@ -175,10 +196,11 @@ class Record this->data_ = data; this->len_ = len; } + void set_data_owner(char *data, int len) { ASSERT(len != 0, "the len of data should not be 0"); - this->~Record(); + reset(); this->data_ = data; this->len_ = len; @@ -211,18 +233,60 @@ class Record return RC::SUCCESS; } - RC set_field(int field_offset, int field_len, char *data) + RC set_field(int field_offset, int field_len, const Value &value) { - if (!owner_) { - LOG_ERROR("cannot set field when record does not own the memory"); - return RC::INTERNAL; - } + // 只警告不检查试试看 + // if (!owner_) { + // LOG_WARN("cannot set field when record does not own the memory"); + // return RC::INTERNAL; + // } if (field_offset + field_len > len_) { LOG_ERROR("invalid offset or length. offset=%d, length=%d, total length=%d", field_offset, field_len, len_); return RC::INVALID_ARGUMENT; } + // 实际数据长度 + auto len = std::min(field_len, value.length()); + // 如果是字符串类型，长度可变，要根据实际长度拷贝数据 + memcpy(data_ + field_offset, value.data(), len); + // 因为列数据是连续的，如果中间某些列加了'\0'，会导致后面列没数据 + // 需要判断更新的字符串是否小于上限，只有小于才需要加'\0'，而大于应该抛出异常 + // 除了字符串类型其他都是定长的 + if (len < field_len) { + data_[field_offset + len] = '\0'; + } + return RC::SUCCESS; + } + + RC get_field(const FieldMeta &field_meta, Value &value) const + { + int field_offset = field_meta.offset(); + int data_len = field_meta.len() - field_meta.nullable(); + + if (field_offset + field_meta.len() > len_) { + LOG_ERROR("invalid offset or length. offset=%d, length=%d, total length=%d", field_offset, field_meta.len(), len_); + return RC::INVALID_ARGUMENT; + } + + value.set_type(field_meta.type()); + + if (field_meta.nullable()) { + // 只有字段是可为空的，取标记位才有意义 + bool is_null = data_[field_offset + field_meta.len() - 1] == '1'; + if (is_null) { + value.set_null(); + return RC::SUCCESS; + } + } + + char *data = new char[data_len]; + memcpy(data, data_ + field_offset, data_len); + value.set_data(data, data_len); + + // vector 不释放内存 + if (!(field_meta.type() == AttrType::VECTORS)) { + delete[] data; + } - memcpy(data_ + field_offset, data, field_len); return RC::SUCCESS; } @@ -236,13 +300,23 @@ class Record this->rid_.page_num = page_num; this->rid_.slot_num = slot_num; } + RID &rid() { return rid_; } const RID &rid() const { return rid_; } + void append_base_rid(BaseTable *table, RID rid) { base_rids_.emplace_back(table, rid); } + + void set_base_rids(std::vector> &base_rids) { base_rids_ = std::move(base_rids); } + + std::vector> &base_rids() { return base_rids_; } + + const std::vector> &base_rids() const { return base_rids_; } + private: - RID rid_; + RID rid_; // 存储基表的记录位置 + std::vector> base_rids_; // 用于视图存储当前记录由哪些基表的哪些记录组成 char *data_ = nullptr; int len_ = 0; /// 如果不是record自己来管理内存，这个字段可能是无效的 - bool owner_ = false; /// 表示当前是否由record来管理内存 + bool owner_ = false; /// 表示当前是否由record来管理内 }; diff --git a/src/observer/storage/record/record_manager.cpp b/src/observer/storage/record/record_manager.cpp index 00f403d9..2fb68f6a 100644 --- a/src/observer/storage/record/record_manager.cpp +++ b/src/observer/storage/record/record_manager.cpp @@ -91,6 +91,14 @@ RC RecordPageIterator::next(Record &record) return record.rid().slot_num != -1 ? RC::SUCCESS : RC::RECORD_EOF; } +RC RecordPageIterator::cleanup() +{ + record_page_handler_ = nullptr; + page_num_ = BP_INVALID_PAGE_NUM; + next_slot_num_ = 0; + return RC::SUCCESS; +} + //////////////////////////////////////////////////////////////////////////////// RecordPageHandler::~RecordPageHandler() { cleanup(); } @@ -352,7 +360,7 @@ RC RowRecordPageHandler::delete_record(const RID *rid) RC RowRecordPageHandler::update_record(const RID &rid, const char *data) { - ASSERT(rw_mode_ != ReadWriteMode::READ_ONLY, "cannot delete record from page while the page is readonly"); + ASSERT(rw_mode_ != ReadWriteMode::READ_ONLY, "cannot update record from page while the page is readonly"); if (rid.slot_num >= page_header_->record_capacity) { LOG_ERROR("Invalid slot_num %d, exceed page's record capacity, frame=%s, page_header=%s", @@ -648,6 +656,31 @@ RC RecordFileHandler::delete_record(const RID *rid) return rc; } +RC RecordFileHandler::update_record(const char *data, const RID *rid) +{ + RC rc = RC::SUCCESS; + + unique_ptr record_page_handler(RecordPageHandler::create(storage_format_)); + + rc = record_page_handler->init(*disk_buffer_pool_, *log_handler_, rid->page_num, ReadWriteMode::READ_WRITE); + if (OB_FAIL(rc)) { + LOG_ERROR("Failed to init record page handler.page number=%d. rc=%s", rid->page_num, strrc(rc)); + return rc; + } + + rc = record_page_handler->update_record(*rid, data); + // 📢 这里注意要清理掉资源，否则会与insert_record中的加锁顺序冲突而可能出现死锁 + // delete record的加锁逻辑是拿到页面锁，删除指定记录，然后加上和释放record manager锁 + // insert record是加上 record manager锁，然后拿到指定页面锁再释放record manager锁 + record_page_handler->cleanup(); + if (OB_SUCC(rc)) { + // 因为这里已经释放了页面锁，并发时，其它线程可能又把该页面填满了，那就不应该再放入 free_pages_ + // 中。但是这里可以不关心，因为在查找空闲页面时，会自动过滤掉已经满的页面 + LOG_TRACE("update record on rid{%d,%d} success", rid->page_num, rid->slot_num); + } + return rc; +} + RC RecordFileHandler::get_record(const RID &rid, Record &record) { unique_ptr page_handler(RecordPageHandler::create(storage_format_)); @@ -670,7 +703,7 @@ RC RecordFileHandler::get_record(const RID &rid, Record &record) return rc; } -RC RecordFileHandler::visit_record(const RID &rid, function updater) +RC RecordFileHandler::visit_record(const RID &rid, const function &updater) { unique_ptr page_handler(RecordPageHandler::create(storage_format_)); @@ -828,6 +861,7 @@ RC RecordFileScanner::close_scan() record_page_handler_->cleanup(); delete record_page_handler_; record_page_handler_ = nullptr; + record_page_iterator_.cleanup(); } return RC::SUCCESS; diff --git a/src/observer/storage/record/record_manager.h b/src/observer/storage/record/record_manager.h index 810ec586..8228464a 100644 --- a/src/observer/storage/record/record_manager.h +++ b/src/observer/storage/record/record_manager.h @@ -111,6 +111,11 @@ class RecordPageIterator */ bool is_valid() const { return record_page_handler_ != nullptr; } + /** + * @brief 操作结束后做的清理工作，初始化成员变量为默认值 + */ + RC cleanup(); + private: RecordPageHandler *record_page_handler_ = nullptr; PageNum page_num_ = BP_INVALID_PAGE_NUM; @@ -357,7 +362,7 @@ class PaxRecordPageHandler : public RecordPageHandler class RecordFileHandler { public: - RecordFileHandler(StorageFormat storage_format) : storage_format_(storage_format){}; + RecordFileHandler(StorageFormat storage_format) : storage_format_(storage_format) {}; ~RecordFileHandler(); /** @@ -388,6 +393,14 @@ class RecordFileHandler */ RC insert_record(const char *data, int record_size, RID *rid); + /** + * @brief 从指定文件中更新指定槽位的记录 + * + * @param data 记录内容 + * @param rid 待更新记录的标识符 + */ + RC update_record(const char *data, const RID *rid); + /** * @brief 数据库恢复时，在指定文件指定位置插入数据 * @@ -399,7 +412,7 @@ class RecordFileHandler RC get_record(const RID &rid, Record &record); - RC visit_record(const RID &rid, function updater); + RC visit_record(const RID &rid, const function &updater); private: /** diff --git a/src/observer/storage/table/base_table.cpp b/src/observer/storage/table/base_table.cpp new file mode 100644 index 00000000..af764b2c --- /dev/null +++ b/src/observer/storage/table/base_table.cpp @@ -0,0 +1,104 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/17 * + * @Description : table base source file * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "base_table.h" + +RC BaseTable::set_value_to_record(char *record_data, const Value &value, const FieldMeta *field) +{ + size_t copy_len = field->len(); + const size_t data_len = value.length(); + if (field->type() == AttrType::CHARS || field->type() == AttrType::TEXTS) { + if (copy_len > data_len) { + copy_len = data_len + 1; + } + } + if (field->type() == AttrType::VECTORS) { + if (copy_len / sizeof(float) != data_len / sizeof(float)) { + return RC::VECTOR_DIM_MISMATCH; + } + if (copy_len > data_len) { + copy_len = data_len; + } + } + // text 类型的话最多存 65535 字节，超出则报错 + memcpy(record_data + field->offset(), value.data(), copy_len); + return RC::SUCCESS; +} + +RC BaseTable::make_record(int value_num, const Value *values, Record &record) +{ + RC rc = RC::SUCCESS; + // 检查字段类型是否一致 + if (value_num + table_meta_.sys_field_num() != table_meta_.field_num()) { + LOG_WARN("Input values don't match the table's schema, table name:%s", table_meta_.name()); + return RC::SCHEMA_FIELD_MISSING; + } + + const int normal_field_start_index = table_meta_.sys_field_num(); + // 复制所有字段的值 + int record_size = table_meta_.record_size(); + char *record_data = (char *)malloc(record_size); + memset(record_data, 0, record_size); + + for (int i = 0; i < value_num && OB_SUCC(rc); i++) { + const FieldMeta *field = table_meta_.field(i + normal_field_start_index); + const Value &value = values[i]; + // 判断是否在 NOT NULL 字段设置 NULL 值 + if (value.is_null()) { + if (!field->nullable()) { + return RC::NOT_NULLABLE_VALUE; + } + record_data[field->offset() + field->len() - 1] = '1'; + } else { + Value real_value = value; + if (field->type() != value.attr_type()) { + if (field->type() == AttrType::TEXTS && value.attr_type() == AttrType::CHARS) { + // 对于超长文本通过借用的方法减少拷贝 + rc = real_value.borrow_text(value); + if (OB_FAIL(rc)) { + LOG_WARN("failed to borrow text value. table name:%s, field name:%s, value length:%d", + table_meta_.name(), field->name(), value.length()); + break; + } + } else { + // 插入不允许非目标类型的类型提升 + rc = Value::cast_to(value, field->type(), real_value, false); + if (OB_FAIL(rc)) { + LOG_WARN("failed to cast value. table name:%s, field name:%s, value:%s", + table_meta_.name(), field->name(), value.to_string().c_str()); + break; + } + } + } + // 进行长度校验 + if (real_value.length() > field->len() - field->nullable()) { + LOG_ERROR("Value length exceeds maximum allowed length for field. Field: %s, Type: %s, Offset: %d, Length: %d, Max Length: %d", + field->name(), + attr_type_to_string(field->type()), + field->offset(), + value.length(), + field->len()); + return RC::VALUE_TOO_LONG; + } + rc = set_value_to_record(record_data, real_value, field); + } + } + + if (OB_FAIL(rc)) { + LOG_WARN("failed to make record. table name:%s", table_meta_.name()); + free(record_data); + return rc; + } + + record.set_data_owner(record_data, record_size); + return RC::SUCCESS; +} diff --git a/src/observer/storage/table/base_table.h b/src/observer/storage/table/base_table.h new file mode 100644 index 00000000..1673a4b7 --- /dev/null +++ b/src/observer/storage/table/base_table.h @@ -0,0 +1,75 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : table base class * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "storage/record/record.h" +#include "storage/table/table_meta.h" +#include "storage/buffer/disk_buffer_pool.h" + +class Db; + +class BaseTable +{ +public: + virtual ~BaseTable() = default; + + Db *db() const { return db_; } + TableType type() const { return table_meta_.table_type(); } + int32_t table_id() const { return table_meta_.table_id(); } + const TableMeta &table_meta() const { return table_meta_; } + const char *name() const { return table_meta_.name(); } + bool is_mutable() const { return table_meta_.is_mutable(); } + + /** + * @brief 根据给定的字段生成一个记录/行 + * @details 通常是由用户传过来的字段，按照schema信息组装成一个record。 + * @param value_num 字段的个数 + * @param values 每个字段的值 + * @param record 生成的记录数据 + */ + RC make_record(int value_num, const Value *values, Record &record); + + RC set_value_to_record(char *record_data, const Value &value, const FieldMeta *field); + + /** + * 打开一个表或视图 + * @param meta_file 保存表元数据的文件完整路径 + * @param base_dir 表所在的文件夹，表记录数据文件、索引数据文件存放位置 + */ + virtual RC open(Db *db, const char *meta_file, const char *base_dir) = 0; + + /** + * @brief 可以在页面锁保护的情况下访问记录 + * @details 当前是在事务中访问记录，为了提供一个“原子性”的访问模式 + * @param rid + * @param visitor + * @return RC + */ + virtual RC visit_record(const RID &rid, const function &visitor) = 0; + + virtual RC insert_record(Record &record) = 0; + virtual RC delete_record(const Record &record) = 0; + virtual RC delete_record(const RID &rid) = 0; + virtual RC update_record(const Record &old_record, const Record &new_record) = 0; + virtual RC get_record(const RID &rid, Record &record) = 0; + + virtual RC drop() = 0; + + virtual RC sync() = 0; + +protected: + Db *db_ = nullptr; + string base_dir_; + TableMeta table_meta_{}; + DiskBufferPool *data_buffer_pool_ = nullptr; /// 数据文件关联的buffer pool +}; diff --git a/src/observer/storage/table/table.cpp b/src/observer/storage/table/table.cpp index 5396c9f1..c457158a 100644 --- a/src/observer/storage/table/table.cpp +++ b/src/observer/storage/table/table.cpp @@ -15,6 +15,8 @@ See the Mulan PSL v2 for more details. */ #include #include +#include + #include "common/defs.h" #include "common/lang/string.h" #include "common/lang/span.h" @@ -30,6 +32,9 @@ See the Mulan PSL v2 for more details. */ #include "storage/record/record_manager.h" #include "storage/table/table.h" #include "storage/trx/trx.h" +#include "storage/db/db.h" +#include "storage/index/ivfflat_index.h" +#include "sql/expr/expression.h" Table::~Table() { @@ -70,6 +75,13 @@ RC Table::create(Db *db, int32_t table_id, const char *path, const char *name, c LOG_WARN("Invalid arguments. table_name=%s, attribute_count=%d", name, attributes.size()); return RC::INVALID_ARGUMENT; } + for (const auto &att : attributes) { + if (att.type == AttrType::VECTORS) { + if (att.length > 16000 * sizeof(float) + 1) { + return RC::INVALID_ARGUMENT; + } + } + } RC rc = RC::SUCCESS; @@ -89,7 +101,8 @@ RC Table::create(Db *db, int32_t table_id, const char *path, const char *name, c // 创建文件 const vector *trx_fields = db->trx_kit().trx_fields(); - if ((rc = table_meta_.init(table_id, name, trx_fields, attributes, storage_format)) != RC::SUCCESS) { + if ((rc = table_meta_.init(table_id, TableType::Table, true, name, trx_fields, attributes, storage_format)) != + RC::SUCCESS) { LOG_ERROR("Failed to init table meta. name:%s, ret:%d", name, rc); return rc; // delete table file } @@ -127,6 +140,41 @@ RC Table::create(Db *db, int32_t table_id, const char *path, const char *name, c return rc; } +RC Table::drop() +{ + auto rc = sync(); // 刷新所有脏页 + if (rc != RC::SUCCESS) { + return rc; + } + + auto table_name = name(); + error_code ec; + auto path = table_meta_file(base_dir_.c_str(), table_name); + if (!filesystem::remove(path, ec)) { + LOG_ERROR("Drop table meta fail: %s. error=%s", path.c_str(), strerror(errno)); + return RC::IOERR_WRITE; + } + + path = table_data_file(base_dir_.c_str(), table_name); + if (!filesystem::remove(path, ec)) { + LOG_ERROR("Drop table data fail: %s. error=%s", path.c_str(), strerror(errno)); + return RC::IOERR_WRITE; + } + + auto index_num = table_meta_.index_num(); + for (int i = 0; i < index_num; ++i) { + ((BplusTreeIndex *)indexes_[i])->close(); + auto index_name = table_meta_.index(i)->name(); + path = table_index_file(base_dir_.c_str(), table_name, index_name); + if (!filesystem::remove(path, ec)) { + LOG_ERROR("Drop table index data fail: %s. error=%s", path.c_str(), strerror(errno)); + return RC::IOERR_WRITE; + } + } + + return RC::SUCCESS; +} + RC Table::open(Db *db, const char *meta_file, const char *base_dir) { // 加载元数据文件 @@ -158,19 +206,11 @@ RC Table::open(Db *db, const char *meta_file, const char *base_dir) const int index_num = table_meta_.index_num(); for (int i = 0; i < index_num; i++) { const IndexMeta *index_meta = table_meta_.index(i); - const FieldMeta *field_meta = table_meta_.field(index_meta->field()); - if (field_meta == nullptr) { - LOG_ERROR("Found invalid index meta info which has a non-exists field. table=%s, index=%s, field=%s", - name(), index_meta->name(), index_meta->field()); - // skip cleanup - // do all cleanup action in destructive Table function - return RC::INTERNAL; - } BplusTreeIndex *index = new BplusTreeIndex(); string index_file = table_index_file(base_dir, name(), index_meta->name()); - rc = index->open(this, index_file.c_str(), *index_meta, *field_meta); + rc = index->open(this, index_file.c_str(), *index_meta); if (rc != RC::SUCCESS) { delete index; LOG_ERROR("Failed to open index. table=%s, index=%s, file=%s, rc=%s", @@ -210,7 +250,7 @@ RC Table::insert_record(Record &record) return rc; } -RC Table::visit_record(const RID &rid, function visitor) +RC Table::visit_record(const RID &rid, const function &visitor) { return record_handler_->visit_record(rid, visitor); } @@ -251,64 +291,6 @@ RC Table::recover_insert_record(Record &record) return rc; } -const char *Table::name() const { return table_meta_.name(); } - -const TableMeta &Table::table_meta() const { return table_meta_; } - -RC Table::make_record(int value_num, const Value *values, Record &record) -{ - RC rc = RC::SUCCESS; - // 检查字段类型是否一致 - if (value_num + table_meta_.sys_field_num() != table_meta_.field_num()) { - LOG_WARN("Input values don't match the table's schema, table name:%s", table_meta_.name()); - return RC::SCHEMA_FIELD_MISSING; - } - - const int normal_field_start_index = table_meta_.sys_field_num(); - // 复制所有字段的值 - int record_size = table_meta_.record_size(); - char *record_data = (char *)malloc(record_size); - memset(record_data, 0, record_size); - - for (int i = 0; i < value_num && OB_SUCC(rc); i++) { - const FieldMeta *field = table_meta_.field(i + normal_field_start_index); - const Value & value = values[i]; - if (field->type() != value.attr_type()) { - Value real_value; - rc = Value::cast_to(value, field->type(), real_value); - if (OB_FAIL(rc)) { - LOG_WARN("failed to cast value. table name:%s,field name:%s,value:%s ", - table_meta_.name(), field->name(), value.to_string().c_str()); - break; - } - rc = set_value_to_record(record_data, real_value, field); - } else { - rc = set_value_to_record(record_data, value, field); - } - } - if (OB_FAIL(rc)) { - LOG_WARN("failed to make record. table name:%s", table_meta_.name()); - free(record_data); - return rc; - } - - record.set_data_owner(record_data, record_size); - return RC::SUCCESS; -} - -RC Table::set_value_to_record(char *record_data, const Value &value, const FieldMeta *field) -{ - size_t copy_len = field->len(); - const size_t data_len = value.length(); - if (field->type() == AttrType::CHARS) { - if (copy_len > data_len) { - copy_len = data_len + 1; - } - } - memcpy(record_data + field->offset(), value.data(), copy_len); - return RC::SUCCESS; -} - RC Table::init_record_handler(const char *base_dir) { string data_file = table_data_file(base_dir, table_meta_.name()); @@ -353,19 +335,20 @@ RC Table::get_chunk_scanner(ChunkFileScanner &scanner, Trx *trx, ReadWriteMode m return rc; } -RC Table::create_index(Trx *trx, const FieldMeta *field_meta, const char *index_name) +RC Table::create_index( + Trx *trx, IndexType index_type, const vector &field_meta, const char *index_name, bool unique) { - if (common::is_blank(index_name) || nullptr == field_meta) { + if (common::is_blank(index_name)) { LOG_INFO("Invalid input arguments, table name is %s, index_name is blank or attribute_name is blank", name()); return RC::INVALID_ARGUMENT; } IndexMeta new_index_meta; - RC rc = new_index_meta.init(index_name, *field_meta); + RC rc = new_index_meta.init(index_name, index_type, field_meta, unique); if (rc != RC::SUCCESS) { - LOG_INFO("Failed to init IndexMeta in table:%s, index_name:%s, field_name:%s", - name(), index_name, field_meta->name()); + LOG_INFO("Failed to init IndexMeta in table:%s, index:%s", + name(), new_index_meta.to_string().c_str()); return rc; } @@ -373,7 +356,7 @@ RC Table::create_index(Trx *trx, const FieldMeta *field_meta, const char *index_ BplusTreeIndex *index = new BplusTreeIndex(); string index_file = table_index_file(base_dir_.c_str(), name(), index_name); - rc = index->create(this, index_file.c_str(), new_index_meta, *field_meta); + rc = index->create(this, index_file.c_str(), new_index_meta); if (rc != RC::SUCCESS) { delete index; LOG_ERROR("Failed to create bplus tree index. file name=%s, rc=%d:%s", index_file.c_str(), rc, strrc(rc)); @@ -451,6 +434,113 @@ RC Table::create_index(Trx *trx, const FieldMeta *field_meta, const char *index_ return rc; } +RC Table::create_vector_index(Trx *trx, IndexType index_type, const vector &field_meta, + const char *index_name, NormalFunctionType distance_type, const std::vector &options) +{ + if (common::is_blank(index_name)) { + LOG_INFO("Invalid input arguments, table name is %s, index_name is blank or attribute_name is blank", name()); + return RC::INVALID_ARGUMENT; + } + + IndexMeta new_index_meta; + + RC rc = new_index_meta.init(index_name, index_type, field_meta); + if (rc != RC::SUCCESS) { + LOG_INFO("Failed to init IndexMeta in table:%s, index:%s", + name(), new_index_meta.to_string().c_str()); + return rc; + } + + // 创建索引相关数据 + auto index = new IvfflatIndex(); + string index_file = table_index_file(base_dir_.c_str(), name(), index_name); + rc = index->create(this, index_file.c_str(), new_index_meta, field_meta[0]); + if (rc != RC::SUCCESS) { + delete index; + LOG_ERROR("Failed to create Ivfflat index. file name=%s, rc=%d:%s", index_file.c_str(), rc, strrc(rc)); + return rc; + } + + // 遍历当前的所有数据，插入这个索引 + RecordFileScanner scanner; + rc = get_record_scanner(scanner, trx, ReadWriteMode::READ_ONLY); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to create scanner while creating vector index. table=%s, index=%s, rc=%s", + name(), index_name, strrc(rc)); + return rc; + } + + // 一次性把某向量类型列数据都读出来 + Record record; + std::vector, RID>> data; + while (OB_SUCC(rc = scanner.next(record))) { + Value value; + // 目前向量仅在一列上建立索引 + rc = record.get_field(field_meta[0], value); + if (OB_FAIL(rc)) { + return rc; + } + data.emplace_back(value.get_vector(), record.rid()); + } + + if (RC::RECORD_EOF == rc) { + rc = RC::SUCCESS; + } else { + LOG_WARN("failed to get record while creating index. table=%s, index=%s, rc=%s", + name(), index_name, strrc(rc)); + return rc; + } + + scanner.close_scan(); + + // 建立向量索引 + index->build_index(data, distance_type, options); + + LOG_INFO("inserted all records into new index. table=%s, index=%s", name(), index_name); + + indexes_.emplace_back(index); + + /// 接下来将这个索引放到表的元数据中 + TableMeta new_table_meta(table_meta_); + rc = new_table_meta.add_index(new_index_meta); + if (rc != RC::SUCCESS) { + LOG_ERROR("Failed to add index (%s) on table (%s). error=%d:%s", index_name, name(), rc, strrc(rc)); + return rc; + } + + /// 内存中有一份元数据，磁盘文件也有一份元数据。修改磁盘文件时，先创建一个临时文件，写入完成后再rename为正式文件 + /// 这样可以防止文件内容不完整 + // 创建元数据临时文件 + string tmp_file = table_meta_file(base_dir_.c_str(), name()) + ".tmp"; + fstream fs; + fs.open(tmp_file, ios_base::out | ios_base::binary | ios_base::trunc); + if (!fs.is_open()) { + LOG_ERROR("Failed to open file for write. file name=%s, errmsg=%s", tmp_file.c_str(), strerror(errno)); + return RC::IOERR_OPEN; // 创建索引中途出错，要做还原操作 + } + if (new_table_meta.serialize(fs) < 0) { + LOG_ERROR("Failed to dump new table meta to file: %s. sys err=%d:%s", tmp_file.c_str(), errno, strerror(errno)); + return RC::IOERR_WRITE; + } + fs.close(); + + // 覆盖原始元数据文件 + string meta_file = table_meta_file(base_dir_.c_str(), name()); + + int ret = rename(tmp_file.c_str(), meta_file.c_str()); + if (ret != 0) { + LOG_ERROR("Failed to rename tmp meta file (%s) to normal meta file (%s) while creating index (%s) on table (%s). " + "system error=%d:%s", + tmp_file.c_str(), meta_file.c_str(), index_name, name(), errno, strerror(errno)); + return RC::IOERR_WRITE; + } + + table_meta_.swap(new_table_meta); + + LOG_INFO("Successfully added a new index (%s) on the table (%s)", index_name, name()); + return rc; +} + RC Table::delete_record(const RID &rid) { RC rc = RC::SUCCESS; @@ -476,6 +566,35 @@ RC Table::delete_record(const Record &record) return rc; } +RC Table::update_record(const Record &old_record, const Record &new_record) +{ + RC rc = RC::SUCCESS; + // 维护索引，先删除后插入 + for (Index *index : indexes_) { + rc = index->delete_entry(old_record.data(), &old_record.rid()); + ASSERT(RC::SUCCESS == rc, + "failed to delete entry from index. table name=%s, index name=%s, rid=%s, rc=%s", + name(), index->index_meta().name(), old_record.rid().to_string().c_str(), strrc(rc)); + } + + // 尝试插入 + rc = insert_entry_of_indexes(new_record.data(), new_record.rid()); + // 出现重复键 + if (rc != RC::SUCCESS) { + // 因为有些索引还没有插入，删除失败不应该报错 + RC delete_entry_of_indexes_rc = delete_entry_of_indexes(new_record.data(), new_record.rid(), false); + if (RC::SUCCESS != delete_entry_of_indexes_rc) { + LOG_WARN("failed to rollback index data when insert index entries failed. table name=%s, rc=%s", name(), strrc(delete_entry_of_indexes_rc)); + return delete_entry_of_indexes_rc; + } + return rc; + } + + // 最后更新记录 + rc = record_handler_->update_record(new_record.data(), &new_record.rid()); + return rc; +} + RC Table::insert_entry_of_indexes(const char *record, const RID &rid) { RC rc = RC::SUCCESS; @@ -511,12 +630,33 @@ Index *Table::find_index(const char *index_name) const } return nullptr; } + Index *Table::find_index_by_field(const char *field_name) const { - const TableMeta &table_meta = this->table_meta(); - const IndexMeta *index_meta = table_meta.find_index_by_field(field_name); - if (index_meta != nullptr) { - return this->find_index(index_meta->name()); + for (const auto &index : indexes_) { + if (index->index_meta().fields().size() == 1) { + auto name = index->index_meta().fields().front().name(); + if (0 == strcmp(name, field_name)) { + return index; + } + } + } + return nullptr; +} + +// 对向量距离类型，索引字段进行检查 +Index *Table::find_vector_index(NormalFunctionType distance_fn, const char *field_name) const +{ + for (const auto &index : indexes_) { + if (index->is_vector_index()) { + auto vector_index = dynamic_cast(index); + if (vector_index->distance_fn() == distance_fn) { + auto name = index->index_meta().fields().front().name(); + if (0 == strcmp(name, field_name)) { + return index; + } + } + } } return nullptr; } diff --git a/src/observer/storage/table/table.h b/src/observer/storage/table/table.h index 089e13ae..a9feb4ce 100644 --- a/src/observer/storage/table/table.h +++ b/src/observer/storage/table/table.h @@ -14,10 +14,10 @@ See the Mulan PSL v2 for more details. */ #pragma once -#include "storage/table/table_meta.h" +#include "storage/table/base_table.h" #include "common/types.h" #include "common/lang/span.h" -#include "common/lang/functional.h" +#include "sql/builtin/builtin.h" struct RID; class Record; @@ -31,17 +31,16 @@ class Index; class IndexScanner; class RecordDeleter; class Trx; -class Db; /** * @brief 表 * */ -class Table +class Table : public BaseTable { public: Table() = default; - ~Table(); + ~Table() override; /** * 创建一个表 @@ -55,35 +54,35 @@ class Table span attributes, StorageFormat storage_format); /** - * 打开一个表 - * @param meta_file 保存表元数据的文件完整路径 - * @param base_dir 表所在的文件夹，表记录数据文件、索引数据文件存放位置 + * 删除一个表 */ - RC open(Db *db, const char *meta_file, const char *base_dir); + RC drop() override; /** - * @brief 根据给定的字段生成一个记录/行 - * @details 通常是由用户传过来的字段，按照schema信息组装成一个record。 - * @param value_num 字段的个数 - * @param values 每个字段的值 - * @param record 生成的记录数据 + * 打开一个表 + * @param meta_file 保存表元数据的文件完整路径 + * @param base_dir 表所在的文件夹，表记录数据文件、索引数据文件存放位置 */ - RC make_record(int value_num, const Value *values, Record &record); + RC open(Db *db, const char *meta_file, const char *base_dir) override; /** * @brief 在当前的表中插入一条记录 * @details 在表文件和索引中插入关联数据。这里只管在表中插入数据，不关心事务相关操作。 * @param record[in/out] 传入的数据包含具体的数据，插入成功会通过此字段返回RID */ - RC insert_record(Record &record); - RC delete_record(const Record &record); - RC delete_record(const RID &rid); - RC get_record(const RID &rid, Record &record); + RC insert_record(Record &record) override; + RC delete_record(const Record &record) override; + RC delete_record(const RID &rid) override; + RC update_record(const Record &old_record, const Record &new_record) override; + RC get_record(const RID &rid, Record &record) override; RC recover_insert_record(Record &record); - // TODO refactor - RC create_index(Trx *trx, const FieldMeta *field_meta, const char *index_name); + RC create_index( + Trx *trx, IndexType index_type, const vector &field_meta, const char *index_name, bool unique); + + RC create_vector_index(Trx *trx, IndexType index_type, const vector &field_meta, const char *index_name, + NormalFunctionType distance_type, const std::vector &options); RC get_record_scanner(RecordFileScanner &scanner, Trx *trx, ReadWriteMode mode); @@ -98,22 +97,16 @@ class Table * @param visitor * @return RC */ - RC visit_record(const RID &rid, function visitor); + RC visit_record(const RID &rid, const function &visitor) override; public: - int32_t table_id() const { return table_meta_.table_id(); } - const char *name() const; - Db *db() const { return db_; } - const TableMeta &table_meta() const; - - RC sync(); + RC sync() override; private: RC insert_entry_of_indexes(const char *record, const RID &rid); RC delete_entry_of_indexes(const char *record, const RID &rid, bool error_on_not_exists); - RC set_value_to_record(char *record_data, const Value &value, const FieldMeta *field); private: RC init_record_handler(const char *base_dir); @@ -121,12 +114,9 @@ class Table public: Index *find_index(const char *index_name) const; Index *find_index_by_field(const char *field_name) const; + Index *find_vector_index(NormalFunctionType distance_fn, const char *field_name) const; private: - Db *db_ = nullptr; - string base_dir_; - TableMeta table_meta_; - DiskBufferPool *data_buffer_pool_ = nullptr; /// 数据文件关联的buffer pool - RecordFileHandler *record_handler_ = nullptr; /// 记录操作 + RecordFileHandler *record_handler_ = nullptr; /// 记录操作 vector indexes_; }; diff --git a/src/observer/storage/table/table_meta.cpp b/src/observer/storage/table/table_meta.cpp index f87cb894..bd130b61 100644 --- a/src/observer/storage/table/table_meta.cpp +++ b/src/observer/storage/table/table_meta.cpp @@ -21,6 +21,8 @@ See the Mulan PSL v2 for more details. */ #include "json/json.h" static const Json::StaticString FIELD_TABLE_ID("table_id"); +static const Json::StaticString FIELD_TABLE_TYPE("table_type"); +static const Json::StaticString FIELD_TABLE_MUTABLE("table_mutable"); static const Json::StaticString FIELD_TABLE_NAME("table_name"); static const Json::StaticString FIELD_STORAGE_FORMAT("storage_format"); static const Json::StaticString FIELD_FIELDS("fields"); @@ -28,6 +30,8 @@ static const Json::StaticString FIELD_INDEXES("indexes"); TableMeta::TableMeta(const TableMeta &other) : table_id_(other.table_id_), + table_type_(other.table_type_), + mutable_(other.mutable_), name_(other.name_), fields_(other.fields_), indexes_(other.indexes_), @@ -43,8 +47,8 @@ void TableMeta::swap(TableMeta &other) noexcept std::swap(record_size_, other.record_size_); } -RC TableMeta::init(int32_t table_id, const char *name, const std::vector *trx_fields, - span attributes, StorageFormat storage_format) +RC TableMeta::init(int32_t table_id, TableType table_type, bool is_mutable, const char *name, + const std::vector *trx_fields, span attributes, StorageFormat storage_format) { if (common::is_blank(name)) { LOG_ERROR("Name cannot be empty"); @@ -67,7 +71,14 @@ RC TableMeta::init(int32_t table_id, const char *name, const std::vectorsize()); for (size_t i = 0; i < trx_fields->size(); i++) { const FieldMeta &field_meta = (*trx_fields)[i]; - fields_[i] = FieldMeta(field_meta.name(), field_meta.type(), field_offset, field_meta.len(), false /*visible*/, field_meta.field_id()); + fields_[i] = FieldMeta(field_meta.name(), + field_meta.type(), + field_offset, + field_meta.len(), + false /*visible*/, + field_meta.field_id(), + field_meta.nullable(), + field_meta.is_mutable()); field_offset += field_meta.len(); } @@ -79,8 +90,14 @@ RC TableMeta::init(int32_t table_id, const char *name, const std::vector TableMeta::trx_fields() const -{ - return span(fields_.data(), sys_field_num()); -} +span TableMeta::trx_fields() const { return span(fields_.data(), sys_field_num()); } const FieldMeta *TableMeta::field(int index) const { return &fields_[index]; } + const FieldMeta *TableMeta::field(const char *name) const { if (nullptr == name) { @@ -127,6 +144,24 @@ const FieldMeta *TableMeta::field(const char *name) const return nullptr; } +RC TableMeta::get_field_metas(const vector &fields, vector &result) const +{ + for (const auto &attribute_name : fields) { + FieldMeta field_meta; + for (const FieldMeta &field : fields_) { + if (0 == strcmp(field.name(), attribute_name.c_str())) { + field_meta = field; + break; + } + } + if (field_meta.len() == 0) { + return RC::SCHEMA_FIELD_NOT_EXIST; + } + result.emplace_back(field_meta); + } + return RC::SUCCESS; +} + const FieldMeta *TableMeta::find_field_by_offset(int offset) const { for (const FieldMeta &field : fields_) { @@ -136,6 +171,7 @@ const FieldMeta *TableMeta::find_field_by_offset(int offset) const } return nullptr; } + int TableMeta::field_num() const { return fields_.size(); } int TableMeta::sys_field_num() const { return static_cast(trx_fields_.size()); } @@ -150,16 +186,6 @@ const IndexMeta *TableMeta::index(const char *name) const return nullptr; } -const IndexMeta *TableMeta::find_index_by_field(const char *field) const -{ - for (const IndexMeta &index : indexes_) { - if (0 == strcmp(index.field(), field)) { - return &index; - } - } - return nullptr; -} - const IndexMeta *TableMeta::index(int i) const { return &indexes_[i]; } int TableMeta::index_num() const { return indexes_.size(); } @@ -169,8 +195,10 @@ int TableMeta::record_size() const { return record_size_; } int TableMeta::serialize(std::ostream &ss) const { Json::Value table_value; - table_value[FIELD_TABLE_ID] = table_id_; - table_value[FIELD_TABLE_NAME] = name_; + table_value[FIELD_TABLE_ID] = table_id_; + table_value[FIELD_TABLE_TYPE] = static_cast(table_type_); + table_value[FIELD_TABLE_MUTABLE] = mutable_; + table_value[FIELD_TABLE_NAME] = name_; table_value[FIELD_STORAGE_FORMAT] = static_cast(storage_format_); Json::Value fields_value; @@ -221,6 +249,22 @@ int TableMeta::deserialize(std::istream &is) int32_t table_id = table_id_value.asInt(); + const Json::Value &table_type_value = table_value[FIELD_TABLE_TYPE]; + if (!table_type_value.isInt()) { + LOG_ERROR("Invalid table type. json value=%s", table_id_value.toStyledString().c_str()); + return -1; + } + + int32_t table_type = table_type_value.asInt(); + + const Json::Value &table_mutable_value = table_value[FIELD_TABLE_MUTABLE]; + if (!table_mutable_value.isBool()) { + LOG_ERROR("Invalid table mutable. json value=%s", table_id_value.toStyledString().c_str()); + return -1; + } + + bool table_mutable = table_mutable_value.asBool(); + const Json::Value &table_name_value = table_value[FIELD_TABLE_NAME]; if (!table_name_value.isString()) { LOG_ERROR("Invalid table name. json value=%s", table_name_value.toStyledString().c_str()); @@ -261,7 +305,9 @@ int TableMeta::deserialize(std::istream &is) auto comparator = [](const FieldMeta &f1, const FieldMeta &f2) { return f1.offset() < f2.offset(); }; std::sort(fields.begin(), fields.end(), comparator); - table_id_ = table_id; + table_id_ = table_id; + table_type_ = static_cast(table_type); + mutable_ = table_mutable; storage_format_ = static_cast(storage_format); name_.swap(table_name); fields_.swap(fields); @@ -269,7 +315,7 @@ int TableMeta::deserialize(std::istream &is) for (const FieldMeta &field_meta : fields_) { if (!field_meta.visible()) { - trx_fields_.push_back(field_meta); // 字段加上trx标识更好 + trx_fields_.push_back(field_meta); // 字段加上trx标识更好 } } @@ -285,7 +331,7 @@ int TableMeta::deserialize(std::istream &is) IndexMeta &index = indexes[i]; const Json::Value &index_value = indexes_value[i]; - rc = IndexMeta::from_json(*this, index_value, index); + rc = IndexMeta::from_json(index_value, index); if (rc != RC::SUCCESS) { LOG_ERROR("Failed to deserialize table meta. table name=%s", table_name.c_str()); return -1; diff --git a/src/observer/storage/table/table_meta.h b/src/observer/storage/table/table_meta.h index 6f438018..f542457a 100644 --- a/src/observer/storage/table/table_meta.h +++ b/src/observer/storage/table/table_meta.h @@ -24,6 +24,13 @@ See the Mulan PSL v2 for more details. */ #include "storage/field/field_meta.h" #include "storage/index/index_meta.h" +enum class TableType +{ + Unknown, + Table, + View +}; + /** * @brief 表元数据 * @@ -38,17 +45,21 @@ class TableMeta : public common::Serializable void swap(TableMeta &other) noexcept; - RC init(int32_t table_id, const char *name, const std::vector *trx_fields, - std::span attributes, StorageFormat storage_format); + RC init(int32_t table_id, TableType table_type, bool is_mutable, const char *name, + const std::vector *trx_fields, std::span attributes, + StorageFormat storage_format); RC add_index(const IndexMeta &index); public: int32_t table_id() const { return table_id_; } + TableType table_type() const { return table_type_; } + bool is_mutable() const { return mutable_; } const char *name() const; const FieldMeta *trx_field() const; const FieldMeta *field(int index) const; const FieldMeta *field(const char *name) const; + RC get_field_metas(const vector &fields, vector &result) const; const FieldMeta *find_field_by_offset(int offset) const; auto field_metas() const -> const std::vector *{ return &fields_; } auto trx_fields() const -> std::span; @@ -58,7 +69,6 @@ class TableMeta : public common::Serializable int sys_field_num() const; const IndexMeta *index(const char *name) const; - const IndexMeta *find_index_by_field(const char *field) const; const IndexMeta *index(int i) const; int index_num() const; @@ -72,7 +82,10 @@ class TableMeta : public common::Serializable void desc(std::ostream &os) const; protected: - int32_t table_id_ = -1; + int32_t table_id_ = -1; + TableType table_type_ = TableType::Unknown; + bool mutable_ = true; // 当前仅对视图可用，是否是只读视图，即包括聚合函数或 groupby having 语句 + std::string name_; std::vector trx_fields_; std::vector fields_; // 包含sys_fields diff --git a/src/observer/storage/table/view.cpp b/src/observer/storage/table/view.cpp new file mode 100644 index 00000000..ba678fea --- /dev/null +++ b/src/observer/storage/table/view.cpp @@ -0,0 +1,404 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#include "storage/db/db.h" +#include "storage/trx/trx.h" +#include "storage/table/view.h" +#include "storage/common/meta_util.h" +#include "sql/stmt/select_stmt.h" + +RC View::create(Db *db, int32_t table_id, const char *path, const char *name, const char *base_dir, + std::vector attr_names, std::string select_sql, SelectStmt *select_stmt, StorageFormat storage_format) +{ + if (table_id < 0) { + LOG_WARN("invalid table id. table_id=%d, table_name=%s", table_id, name); + return RC::INVALID_ARGUMENT; + } + + if (common::is_blank(name)) { + LOG_WARN("Name cannot be empty"); + return RC::INVALID_ARGUMENT; + } + LOG_INFO("Begin to create table %s:%s", base_dir, name); + + RC rc = RC::SUCCESS; + + tables_ = select_stmt->tables(); + bool is_mutable = true; + + auto &query_exprs = select_stmt->query_expressions(); + std::vector attr_infos(query_exprs.size()); + field_index_.resize(query_exprs.size()); + for (int i = 0; i < select_stmt->query_expressions_size(); ++i) { + auto &query_expr = query_exprs[i]; + AttrInfoSqlNode attr_info; + if (query_expr->type() == ExprType::FIELD) { + auto field_expr = dynamic_cast(query_expr.get()); + + // 建立视图字段到基表字段的索引 + bool find = false; + for (auto &table : tables_) { + auto table_field_meta = table->table_meta().field(field_expr->name()); + // 当前视图字段在这个表 + if (table_field_meta != nullptr) { + field_index_[i] = {table, table_field_meta->field_id()}; + find = true; + break; + } + } + if (!find) { + LOG_ERROR("View field '%s' not found in any base tables", field_expr->name()); + return RC::SCHEMA_FIELD_MISSING; + } + + auto field_meta = field_expr->field().meta(); + attr_info.type = field_meta->type(); + attr_info.name = attr_names.empty() ? (field_expr->has_alias() ? field_expr->alias() : field_meta->name()) + : std::move(attr_names[i]); + attr_info.length = field_meta->len(); + attr_info.nullable = field_meta->nullable(); + attr_info.mutable_ = field_meta->is_mutable(); + } else { + // 表达式字段为无效索引 + field_index_[i] = {nullptr, -1}; + + if (query_expr->type() == ExprType::AGGREGATION) { + is_mutable = false; // 包含聚合函数的是只读视图 + } + + attr_info.type = query_expr->value_type(); + attr_info.name = attr_names.empty() ? (query_expr->has_alias() ? query_expr->alias() : query_expr->name()) + : std::move(attr_names[i]); + // 简单处理，非 field 表达式都是可为空的 + attr_info.length = query_expr->value_length() + 1; + attr_info.nullable = true; + attr_info.mutable_ = false; + } + attr_infos[i] = std::move(attr_info); + } + + if (attr_infos.empty()) { + LOG_WARN("Invalid arguments. table_name=%s, attribute_count=%d", name, attr_infos.size()); + return RC::INVALID_ARGUMENT; + } + + // 包含 groupby 和 having 的是只读视图 + if (!select_stmt->group_by().empty() || select_stmt->having_filter_stmt()) { + is_mutable = false; + } + + // 使用 table_name.table 记录一个表的元数据 + // 判断表文件是否已经存在 + int fd = ::open(path, O_WRONLY | O_CREAT | O_EXCL | O_CLOEXEC, 0600); + if (fd < 0) { + if (EEXIST == errno) { + LOG_ERROR("Failed to create table file, it has been created. %s, EEXIST, %s", path, strerror(errno)); + return RC::SCHEMA_TABLE_EXIST; + } + LOG_ERROR("Create table file failed. filename=%s, errmsg=%d:%s", path, errno, strerror(errno)); + return RC::IOERR_OPEN; + } + + close(fd); + + // 创建文件 + const vector *trx_fields = db->trx_kit().trx_fields(); + if ((rc = table_meta_.init(table_id, TableType::View, is_mutable, name, trx_fields, attr_infos, storage_format)) != + RC::SUCCESS) { + LOG_ERROR("Failed to init table meta. name:%s, ret:%d", name, rc); + return rc; // delete table file + } + + fstream fs; + fs.open(path, ios_base::out | ios_base::binary); + if (!fs.is_open()) { + LOG_ERROR("Failed to open file for write. file name=%s, errmsg=%s", path, strerror(errno)); + return RC::IOERR_OPEN; + } + + // 记录元数据到文件中 + table_meta_.serialize(fs); + fs.close(); + + db_ = db; + base_dir_ = base_dir; + + select_sql_ = std::move(select_sql); + + // 存储数据文件 + string data_file = table_data_file(base_dir, name); + std::ofstream file(data_file, ios_base::out | ios_base::binary); + if (!file.is_open()) { + LOG_ERROR("Failed to open file for write. file name=%s, errmsg=%s", path, strerror(errno)); + return RC::IOERR_OPEN; + } + + // 写入查询 sql，存一条 sql 就行了 + file << select_sql_ << std::endl; + file.close(); + + LOG_INFO("Successfully create table %s:%s", base_dir, name); + return rc; +} + +RC View::insert_record(Record &record) +{ + RC rc = RC::SUCCESS; + + for (auto &table : tables_) { + auto value_num = table->table_meta().field_num(); + std::vector values(value_num); + + // 不涉及的字段默认为 null + for (auto &value : values) { + value.set_null(); + } + + auto field_metas = *table_meta_.field_metas(); + // 遍历视图的字段找到基表的字段 + for (int i = 0; i < field_metas.size(); ++i) { + // 基表字段所在索引 + auto [base_table, idx] = field_index_[i]; + // 是一个表则写入 value + if (strcmp(base_table->name(), table->name()) == 0) { + // 插入值 + Value value; + rc = record.get_field(field_metas[i], value); + if (OB_FAIL(rc)) { + return rc; + } + values[idx] = std::move(value); + } + } + + // 生成真正的 record 并插入基表 + Record real_record; + rc = table->make_record(value_num, values.data(), real_record); + if (OB_FAIL(rc)) { + LOG_WARN("failed to make record. rc=%s", strrc(rc)); + return rc; + } + + rc = table->insert_record(real_record); + if (OB_FAIL(rc)) { + LOG_WARN("failed to insert record into table. rc=%s", strrc(rc)); + return rc; + } + } + + return RC::SUCCESS; +} + +RC View::delete_record(const Record &record) +{ + RC rc = RC::SUCCESS; + for (auto &[table, rid] : record.base_rids()) { + rc = table->delete_record(rid); + if (OB_FAIL(rc)) { + return rc; + } + } + return rc; +} + +RC View::delete_record(const RID &rid) { return RC::UNIMPLEMENTED; } + +RC View::update_record(const Record &old_record, const Record &new_record) +{ + // join 视图支持 update 可能会更新多个表 + RC rc = RC::SUCCESS; + + for (auto &[table, rid] : old_record.base_rids()) { + // 得到基表的记录 + Record base_old_record; + rc = table->get_record(rid, base_old_record); + if (OB_FAIL(rc)) { + return rc; + } + + // 将视图的修改记录映射到基表的记录 + Record base_new_record = base_old_record; + auto &field_metas = *table_meta_.field_metas(); + for (int i = 0; i < field_metas.size(); ++i) { + auto &field_meta = field_metas[i]; + if (field_meta.is_mutable()) { + auto &[base_table, idx] = field_index_[i]; + // 是当前表的字段 + if (strcmp(base_table->name(), table->name()) == 0) { + // 取出视图中的值存入待更新的基表记录 + Value value; + rc = new_record.get_field(field_meta, value); + if (OB_FAIL(rc)) { + return rc; + } + auto base_field_meta = base_table->table_meta().field(idx); + rc = base_new_record.set_field(base_field_meta->offset(), base_field_meta->len(), value); + if (OB_FAIL(rc)) { + return rc; + } + } + } + } + + table->update_record(base_old_record, base_new_record); + } + + return rc; +} + +RC View::get_record(const RID &rid, Record &record) { return RC::SUCCESS; } + +RC View::visit_record(const RID &rid, const function &visitor) { return RC::SUCCESS; } + +RC View::sync() +{ + LOG_INFO("Sync view over. view=%s", name()); + return RC::SUCCESS; +} + +RC View::drop() +{ + auto rc = sync(); + if (rc != RC::SUCCESS) { + return rc; + } + + auto table_name = name(); + error_code ec; + auto path = table_meta_file(base_dir_.c_str(), table_name); + if (!filesystem::remove(path, ec)) { + LOG_ERROR("Drop table meta fail: %s. error=%s", path.c_str(), strerror(errno)); + return RC::IOERR_WRITE; + } + + return RC::SUCCESS; +} + +RC View::open(Db *db, const char *meta_file, const char *base_dir) +{ + // 加载元数据文件 + fstream fs; + string meta_file_path = string(base_dir) + common::FILE_PATH_SPLIT_STR + meta_file; + fs.open(meta_file_path, ios_base::in | ios_base::binary); + if (!fs.is_open()) { + LOG_ERROR("Failed to open meta file for read. file name=%s, errmsg=%s", meta_file_path.c_str(), strerror(errno)); + return RC::IOERR_OPEN; + } + if (table_meta_.deserialize(fs) < 0) { + LOG_ERROR("Failed to deserialize table meta. file name=%s", meta_file_path.c_str()); + fs.close(); + return RC::INTERNAL; + } + fs.close(); + + db_ = db; + base_dir_ = base_dir; + + // 加载数据文件 + RC rc = init_data(); + if (rc != RC::SUCCESS) { + LOG_ERROR("Failed to open table %s due to init record handler failed.", base_dir); + // don't need to remove the data_file + return rc; + } + + return RC::SUCCESS; +} + +RC View::init_data() +{ + // 加载数据文件 + string data_file = table_data_file(base_dir_.c_str(), table_meta_.name()); + + // 先读出来查询 sql，构造查询 stmt + // 然后就可以初始化 tables 和建立字段映射 + + std::ifstream file(data_file); + + if (!file.is_open()) { + LOG_ERROR("Failed to open file for read. file name=%s, errmsg=%s", data_file.c_str(), strerror(errno)); + return RC::IOERR_OPEN; + } + + // 读取一整行，默认是读到空格就停了 + std::getline(file, select_sql_); + + return RC::SUCCESS; +} + +RC View::init_member() +{ + RC rc = RC::SUCCESS; + + ParsedSqlResult parsed_sql_result; + + parse(select_sql_.c_str(), &parsed_sql_result); + if (parsed_sql_result.sql_nodes().empty()) { + LOG_ERROR("Parsing failed: No SQL nodes found"); + return RC::INTERNAL; + } + + if (parsed_sql_result.sql_nodes().size() > 1) { + LOG_WARN("got multi sql commands but only 1 will be handled"); + } + + std::unique_ptr sql_node = std::move(parsed_sql_result.sql_nodes().front()); + if (sql_node->flag == SCF_ERROR) { + LOG_ERROR("Syntax error in SQL"); + return RC::SQL_SYNTAX; + } + if (sql_node->flag != SCF_SELECT) { + LOG_ERROR("Unexpected SQL command type. Expected SELECT, got flag: %d", sql_node->flag); + return RC::INTERNAL; + } + + Stmt *stmt = nullptr; + rc = SelectStmt::create(db_, sql_node->selection, stmt); + if (rc != RC::SUCCESS && rc != RC::UNIMPLEMENTED) { + LOG_WARN("Failed to create select stmt. rc=%d:%s", rc, strrc(rc)); + return rc; + } + + // 初始化成员变量 + auto select_stmt = dynamic_cast(stmt); + tables_ = select_stmt->tables(); + + auto &query_exprs = select_stmt->query_expressions(); + field_index_.resize(query_exprs.size()); + for (int i = 0; i < query_exprs.size(); ++i) { + auto &query_expr = query_exprs[i]; + if (query_expr->type() == ExprType::FIELD) { + auto field_expr = dynamic_cast(query_expr.get()); + + // 建立视图字段到基表字段的索引 + bool find = false; + for (auto &table : tables_) { + auto table_field_meta = table->table_meta().field(field_expr->name()); + // 当前视图字段在这个表 + if (table_field_meta != nullptr) { + field_index_[i] = {table, table_field_meta->field_id()}; + find = true; + break; + } + } + if (!find) { + LOG_ERROR("View field '%s' not found in any base tables", field_expr->name()); + return RC::SCHEMA_FIELD_MISSING; + } + } else { + // 表达式字段为无效索引 + field_index_[i] = {nullptr, -1}; + } + } + + return rc; +} diff --git a/src/observer/storage/table/view.h b/src/observer/storage/table/view.h new file mode 100644 index 00000000..ad9bccf3 --- /dev/null +++ b/src/observer/storage/table/view.h @@ -0,0 +1,68 @@ +/*************************************************************** + * * + * @Author : Koschei * + * @Email : nitianzero@gmail.com * + * @Date : 2024/10/12 * + * @Description : Brief description of the file's purpose * + * * + * Copyright (c) 2024 Koschei * + * All rights reserved. * + * * + ***************************************************************/ + +#pragma once + +#include "storage/table/base_table.h" +#include "sql/operator/physical_operator.h" + +class View : public BaseTable +{ +public: + View() = default; + ~View() override = default; + + /** + * 创建一个表 + * @param path 元数据保存的文件(完整路径) + * @param name 表名 + * @param base_dir 表数据存放的路径 + * @param attr_names 字段 + * @param select_sql 查询 sql + * @param select_stmt 查询 stmt + * @param storage_format 存储格式，与基表一致 + */ + RC create(Db *db, int32_t table_id, const char *path, const char *name, const char *base_dir, + std::vector attr_names, std::string select_sql, SelectStmt *select_stmt, + StorageFormat storage_format); + + RC drop() override; + + /** + * 打开一个视图 + * @param meta_file 保存表元数据的文件完整路径 + * @param base_dir 表所在的文件夹，表记录数据文件、索引数据文件存放位置 + */ + RC open(Db *db, const char *meta_file, const char *base_dir) override; + + RC init_data(); + + RC init_member(); + + RC insert_record(Record &record) override; + RC delete_record(const Record &record) override; + RC delete_record(const RID &rid) override; + RC update_record(const Record &old_record, const Record &new_record) override; + RC get_record(const RID &rid, Record &record) override; + RC visit_record(const RID &rid, const function &visitor) override; + + RC sync() override; + + const std::string &select_sql() { return select_sql_; } + + bool has_join() { return tables_.size() > 1; } + +private: + std::string select_sql_; // 持久化，运行时也只能存解析后的 sql，因为涉及独占资源的移动 + std::vector tables_; // 可能含视图和基表，所以要在所有表都加载好再处理 + std::vector> field_index_; // 视图的 field 对应 哪个物理表和对应的 field idx +}; diff --git a/src/observer/storage/trx/mvcc_trx.cpp b/src/observer/storage/trx/mvcc_trx.cpp index eefaef43..3a78de15 100644 --- a/src/observer/storage/trx/mvcc_trx.cpp +++ b/src/observer/storage/trx/mvcc_trx.cpp @@ -13,6 +13,8 @@ See the Mulan PSL v2 for more details. */ // #include "storage/trx/mvcc_trx.h" + +#include #include "storage/db/db.h" #include "storage/field/field.h" #include "storage/trx/mvcc_trx_log.h" @@ -31,10 +33,21 @@ MvccTrxKit::~MvccTrxKit() RC MvccTrxKit::init() { // 事务使用一些特殊的字段，放到每行记录中，表示行记录的可见性。 - fields_ = vector{ - // field_id in trx fields is invisible. - FieldMeta("__trx_xid_begin", AttrType::INTS, 0 /*attr_offset*/, 4 /*attr_len*/, false /*visible*/, -1/*field_id*/), - FieldMeta("__trx_xid_end", AttrType::INTS, 0 /*attr_offset*/, 4 /*attr_len*/, false /*visible*/, -2/*field_id*/)}; + fields_ = vector{// field_id in trx fields is invisible. + FieldMeta("__trx_xid_begin", + AttrType::INTS, + 0 /*attr_offset*/, + 4 /*attr_len*/, + false /*visible*/, + -1 /*field_id*/, + false), + FieldMeta("__trx_xid_end", + AttrType::INTS, + 0 /*attr_offset*/, + 4 /*attr_len*/, + false /*visible*/, + -2 /*field_id*/, + false)}; LOG_INFO("init mvcc trx kit done."); return RC::SUCCESS; @@ -113,11 +126,10 @@ LogReplayer *MvccTrxKit::create_log_replayer(Db &db, LogHandler &log_handler) //////////////////////////////////////////////////////////////////////////////// -MvccTrx::MvccTrx(MvccTrxKit &kit, LogHandler &log_handler) : trx_kit_(kit), log_handler_(log_handler) -{} +MvccTrx::MvccTrx(MvccTrxKit &kit, LogHandler &log_handler) : trx_kit_(kit), log_handler_(log_handler) {} -MvccTrx::MvccTrx(MvccTrxKit &kit, LogHandler &log_handler, int32_t trx_id) - : trx_kit_(kit), log_handler_(log_handler), trx_id_(trx_id) +MvccTrx::MvccTrx(MvccTrxKit &kit, LogHandler &log_handler, int32_t trx_id) + : trx_kit_(kit), log_handler_(log_handler), trx_id_(trx_id) { started_ = true; recovering_ = true; @@ -125,7 +137,7 @@ MvccTrx::MvccTrx(MvccTrxKit &kit, LogHandler &log_handler, int32_t trx_id) MvccTrx::~MvccTrx() {} -RC MvccTrx::insert_record(Table *table, Record &record) +RC MvccTrx::insert_record(BaseTable *table, Record &record) { Field begin_field; Field end_field; @@ -144,11 +156,11 @@ RC MvccTrx::insert_record(Table *table, Record &record) ASSERT(rc == RC::SUCCESS, "failed to append insert record log. trx id=%d, table id=%d, rid=%s, record len=%d, rc=%s", trx_id_, table->table_id(), record.rid().to_string().c_str(), record.len(), strrc(rc)); - operations_.push_back(Operation(Operation::Type::INSERT, table, record.rid())); + operations_.emplace_back(Operation::Type::INSERT, table, record.rid()); return rc; } -RC MvccTrx::delete_record(Table *table, Record &record) +RC MvccTrx::delete_record(BaseTable *table, Record &record) { Field begin_field; Field end_field; @@ -181,12 +193,38 @@ RC MvccTrx::delete_record(Table *table, Record &record) ASSERT(rc == RC::SUCCESS, "failed to append delete record log. trx id=%d, table id=%d, rid=%s, record len=%d, rc=%s", trx_id_, table->table_id(), record.rid().to_string().c_str(), record.len(), strrc(rc)); - operations_.push_back(Operation(Operation::Type::DELETE, table, record.rid())); + operations_.emplace_back(Operation::Type::DELETE, table, record.rid()); return RC::SUCCESS; } -RC MvccTrx::visit_record(Table *table, Record &record, ReadWriteMode mode) +RC MvccTrx::update_record(BaseTable *table, Record &old_record, Record &new_record) +{ + Field begin_field; + Field end_field; + trx_fields(table, begin_field, end_field); + + // 更新的处理和插入一样，begin 设置负值，但是 end 保持不变，因为可能遇到一个老事务对旧数据的修改 + begin_field.set_int(new_record, -trx_id_); + end_field.set_int(new_record, trx_kit_.max_trx_id()); + + RC rc = table->update_record(old_record, new_record); + if (rc != RC::SUCCESS) { + LOG_WARN("failed to insert record into table. rc=%s", strrc(rc)); + return rc; + } + + // 不实现故障恢复应该不用日志 + // rc = log_handler_.insert_record(trx_id_, table, record.rid()); + // ASSERT(rc == RC::SUCCESS, "failed to append insert record log. trx id=%d, table id=%d, rid=%s, record len=%d, + // rc=%s", + // trx_id_, table->table_id(), record.rid().to_string().c_str(), record.len(), strrc(rc)); + + operations_.emplace_back(Operation::Type::UPDATE, table, old_record.rid(), old_record, new_record); + return rc; +} + +RC MvccTrx::visit_record(BaseTable *table, Record &record, ReadWriteMode mode) { Field begin_field; Field end_field; @@ -248,7 +286,7 @@ RC MvccTrx::visit_record(Table *table, Record &record, ReadWriteMode mode) * @param begin_xid_field 返回处理begin_xid的字段 * @param end_xid_field 返回处理end_xid的字段 */ -void MvccTrx::trx_fields(Table *table, Field &begin_xid_field, Field &end_xid_field) const +void MvccTrx::trx_fields(BaseTable *table, Field &begin_xid_field, Field &end_xid_field) const { const TableMeta &table_meta = table->table_meta(); span trx_fields = table_meta.trx_fields(); @@ -280,15 +318,28 @@ RC MvccTrx::commit() RC MvccTrx::commit_with_trx_id(int32_t commit_xid) { // TODO 原子性提交BUG：这里存在一个很大的问题，不能让其他事务一次性看到当前事务更新到的数据或同时看不到 + + // 当前事务在提交时，会逐个修改之前修改过的行数据，调整版本号。 + // 这造成的问题是，在某个时刻，有些行数据的版本号已经修改了，有些还没有。那可能会存在一个事务，能够看到已经修改完成版本号的行，但是看不到未修改的行。 + // 比如事务A，插入了3条数据，在提交的时候，逐个修改版本号，某个情况下可能会存在下面的场景(假设A的事务ID是90，commit + // id是100)： + // + // record begin xid end xid data + // R1 100 +∞ ... + // R2 100 +∞ ... + // R3 -90 +∞ ... + // 此时有一个新的事务，假设事务号是 110，那么它可以看到记录R1和R2，但是看不到R3，因为R3从记录状态来看，还没有提交。 + RC rc = RC::SUCCESS; started_ = false; for (const Operation &operation : operations_) { switch (operation.type()) { case Operation::Type::INSERT: { - RID rid(operation.page_num(), operation.slot_num()); - Table *table = operation.table(); - Field begin_xid_field, end_xid_field; + RID rid = operation.rid(); + BaseTable *table = operation.table(); + + Field begin_xid_field, end_xid_field; trx_fields(table, begin_xid_field, end_xid_field); auto record_updater = [this, &begin_xid_field, commit_xid](Record &record) -> bool { @@ -308,8 +359,8 @@ RC MvccTrx::commit_with_trx_id(int32_t commit_xid) } break; case Operation::Type::DELETE: { - Table *table = operation.table(); - RID rid(operation.page_num(), operation.slot_num()); + RID rid = operation.rid(); + BaseTable *table = operation.table(); Field begin_xid_field, end_xid_field; trx_fields(table, begin_xid_field, end_xid_field); @@ -329,6 +380,29 @@ RC MvccTrx::commit_with_trx_id(int32_t commit_xid) rid.to_string().c_str(), strrc(rc)); } break; + case Operation::Type::UPDATE: { + RID rid = operation.rid(); + BaseTable *table = operation.table(); + + Field begin_xid_field, end_xid_field; + trx_fields(table, begin_xid_field, end_xid_field); + + auto record_updater = [this, &begin_xid_field, commit_xid](Record &record) -> bool { + LOG_DEBUG("before commit update record. trx id=%d, begin xid=%d, commit xid=%d, lbt=%s", + trx_id_, begin_xid_field.get_int(record), commit_xid, lbt()); + ASSERT(begin_xid_field.get_int(record) == -trx_id_, + "got an invalid record while committing. end xid=%d, this trx id=%d", + begin_xid_field.get_int(record), trx_id_); + + begin_xid_field.set_int(record, commit_xid); + return true; + }; + + rc = operation.table()->visit_record(rid, record_updater); + ASSERT(rc == RC::SUCCESS, "failed to get record while committing. rid=%s, rc=%s", + rid.to_string().c_str(), strrc(rc)); + } break; + default: { ASSERT(false, "unsupported operation. type=%d", static_cast(operation.type())); } @@ -350,12 +424,11 @@ RC MvccTrx::rollback() RC rc = RC::SUCCESS; started_ = false; - for (auto iter = operations_.rbegin(), itend = operations_.rend(); iter != itend; ++iter) { - const Operation &operation = *iter; + for (auto &operation : std::ranges::reverse_view(operations_)) { switch (operation.type()) { case Operation::Type::INSERT: { - RID rid(operation.page_num(), operation.slot_num()); - Table *table = operation.table(); + RID rid = operation.rid(); + BaseTable *table = operation.table(); // 这里也可以不删除，仅仅给数据加个标识位，等垃圾回收器来收割也行 if (recovering_) { @@ -382,8 +455,8 @@ RC MvccTrx::rollback() } break; case Operation::Type::DELETE: { - Table *table = operation.table(); - RID rid(operation.page_num(), operation.slot_num()); + RID rid = operation.rid(); + BaseTable *table = operation.table(); ASSERT(rc == RC::SUCCESS, "failed to get record while rollback. rid=%s, rc=%s", rid.to_string().c_str(), strrc(rc)); @@ -408,6 +481,35 @@ RC MvccTrx::rollback() rid.to_string().c_str(), strrc(rc)); } break; + case Operation::Type::UPDATE: { + RID rid = operation.rid(); + BaseTable *table = operation.table(); + + if (recovering_) { + // 恢复的时候，需要额外判断下当前记录是否还是当前事务拥有。是的话才能删除记录 + Record record; + rc = table->get_record(rid, record); + if (OB_SUCC(rc)) { + Field begin_xid_field, end_xid_field; + trx_fields(table, begin_xid_field, end_xid_field); + if (begin_xid_field.get_int(record) != -trx_id_) { + continue; + } + } else if (RC::RECORD_NOT_EXIST == rc) { + continue; + } else { + LOG_WARN("failed to get record while rollback. table=%s, rid=%s, rc=%s", + table->name(), rid.to_string().c_str(), strrc(rc)); + return rc; + } + } + + // 直接用旧的记录的时间戳覆盖 + rc = table->update_record(operation.updated_record(), operation.old_record()); + ASSERT(rc == RC::SUCCESS, "failed to update record while rollback. rid=%s, rc=%s", + rid.to_string().c_str(), strrc(rc)); + } break; + default: { ASSERT(false, "unsupported operation. type=%d", static_cast(operation.type())); } @@ -423,7 +525,7 @@ RC MvccTrx::rollback() return rc; } -RC find_table(Db *db, const LogEntry &log_entry, Table *&table) +RC find_table(Db *db, const LogEntry &log_entry, BaseTable *&table) { auto *trx_log_header = reinterpret_cast(log_entry.data()); switch (MvccTrxLogOperation(trx_log_header->operation_type).type()) { @@ -445,13 +547,15 @@ RC find_table(Db *db, const LogEntry &log_entry, Table *&table) RC MvccTrx::redo(Db *db, const LogEntry &log_entry) { - auto *trx_log_header = reinterpret_cast(log_entry.data()); - Table *table = nullptr; - RC rc = find_table(db, log_entry, table); + auto *trx_log_header = reinterpret_cast(log_entry.data()); + BaseTable *base_table = nullptr; + RC rc = find_table(db, log_entry, base_table); if (OB_FAIL(rc)) { return rc; } + ASSERT(base_table->type() == TableType::Table, "Only tables support MVCC. The provided base_table is not of type Table."); + Table *table = static_cast
(base_table); switch (MvccTrxLogOperation(trx_log_header->operation_type).type()) { case MvccTrxLogOperation::Type::INSERT_RECORD: { auto *trx_log_record = reinterpret_cast(log_entry.data()); diff --git a/src/observer/storage/trx/mvcc_trx.h b/src/observer/storage/trx/mvcc_trx.h index 4a16ba0a..c54f6aac 100644 --- a/src/observer/storage/trx/mvcc_trx.h +++ b/src/observer/storage/trx/mvcc_trx.h @@ -74,10 +74,11 @@ class MvccTrx : public Trx */ MvccTrx(MvccTrxKit &trx_kit, LogHandler &log_handler); MvccTrx(MvccTrxKit &trx_kit, LogHandler &log_handler, int32_t trx_id); // used for recover - virtual ~MvccTrx(); + ~MvccTrx() override; - RC insert_record(Table *table, Record &record) override; - RC delete_record(Table *table, Record &record) override; + RC insert_record(BaseTable *table, Record &record) override; + RC delete_record(BaseTable *table, Record &record) override; + RC update_record(BaseTable *table, Record &old_record, Record &new_record) override; /** * @brief 当访问到某条数据时，使用此函数来判断是否可见，或者是否有访问冲突 @@ -89,7 +90,7 @@ class MvccTrx : public Trx * - RECORD_INVISIBLE 此数据对当前事务不可见，应该跳过 * - LOCKED_CONCURRENCY_CONFLICT 与其它事务有冲突 */ - RC visit_record(Table *table, Record &record, ReadWriteMode mode) override; + RC visit_record(BaseTable *table, Record &record, ReadWriteMode mode) override; RC start_if_need() override; RC commit() override; @@ -101,7 +102,7 @@ class MvccTrx : public Trx private: RC commit_with_trx_id(int32_t commit_id); - void trx_fields(Table *table, Field &begin_xid_field, Field &end_xid_field) const; + void trx_fields(BaseTable *table, Field &begin_xid_field, Field &end_xid_field) const; private: static const int32_t MAX_TRX_ID = numeric_limits::max(); diff --git a/src/observer/storage/trx/mvcc_trx_log.cpp b/src/observer/storage/trx/mvcc_trx_log.cpp index abb4e7ba..dbac4d7d 100644 --- a/src/observer/storage/trx/mvcc_trx_log.cpp +++ b/src/observer/storage/trx/mvcc_trx_log.cpp @@ -68,7 +68,7 @@ MvccTrxLogHandler::MvccTrxLogHandler(LogHandler &log_handler) : log_handler_(log MvccTrxLogHandler::~MvccTrxLogHandler() {} -RC MvccTrxLogHandler::insert_record(int32_t trx_id, Table *table, const RID &rid) +RC MvccTrxLogHandler::insert_record(int32_t trx_id, BaseTable *table, const RID &rid) { ASSERT(trx_id > 0, "invalid trx_id:%d", trx_id); @@ -83,7 +83,7 @@ RC MvccTrxLogHandler::insert_record(int32_t trx_id, Table *table, const RID &rid lsn, LogModule::Id::TRANSACTION, span(reinterpret_cast(&log_entry), sizeof(log_entry))); } -RC MvccTrxLogHandler::delete_record(int32_t trx_id, Table *table, const RID &rid) +RC MvccTrxLogHandler::delete_record(int32_t trx_id, BaseTable *table, const RID &rid) { ASSERT(trx_id > 0, "invalid trx_id:%d", trx_id); @@ -108,7 +108,7 @@ RC MvccTrxLogHandler::commit(int32_t trx_id, int32_t commit_trx_id) log_entry.commit_trx_id = commit_trx_id; LSN lsn = 0; - RC rc = log_handler_.append( + RC rc = log_handler_.append( lsn, LogModule::Id::TRANSACTION, span(reinterpret_cast(&log_entry), sizeof(log_entry))); if (OB_FAIL(rc)) { return rc; @@ -134,7 +134,7 @@ RC MvccTrxLogHandler::rollback(int32_t trx_id) //////////////////////////////////////////////////////////////////////////////////////////////////////////////////////// MvccTrxLogReplayer::MvccTrxLogReplayer(Db &db, MvccTrxKit &trx_kit, LogHandler &log_handler) - : db_(db), trx_kit_(trx_kit), log_handler_(log_handler) + : db_(db), trx_kit_(trx_kit), log_handler_(log_handler) {} RC MvccTrxLogReplayer::replay(const LogEntry &entry) @@ -150,9 +150,9 @@ RC MvccTrxLogReplayer::replay(const LogEntry &entry) return RC::LOG_ENTRY_INVALID; } - auto *header = reinterpret_cast(entry.data()); - MvccTrx *trx = nullptr; - auto trx_iter = trx_map_.find(header->trx_id); + auto *header = reinterpret_cast(entry.data()); + MvccTrx *trx = nullptr; + auto trx_iter = trx_map_.find(header->trx_id); if (trx_iter == trx_map_.end()) { trx = static_cast(trx_kit_.create_trx(log_handler_, header->trx_id)); // trx = new MvccTrx(trx_kit_, log_handler_, header->trx_id); @@ -179,7 +179,7 @@ RC MvccTrxLogReplayer::on_done() /// 日志回放已经完成，需要把没有提交的事务，回滚掉 for (auto &pair : trx_map_) { MvccTrx *trx = pair.second; - trx->rollback(); // 恢复时的rollback，可能遇到之前已经回滚一半的事务又再次调用回滚的情况 + trx->rollback(); // 恢复时的rollback，可能遇到之前已经回滚一半的事务又再次调用回滚的情况 delete pair.second; } trx_map_.clear(); diff --git a/src/observer/storage/trx/mvcc_trx_log.h b/src/observer/storage/trx/mvcc_trx_log.h index 4efc298d..05c04b89 100644 --- a/src/observer/storage/trx/mvcc_trx_log.h +++ b/src/observer/storage/trx/mvcc_trx_log.h @@ -22,7 +22,7 @@ See the Mulan PSL v2 for more details. */ #include "storage/clog/log_replayer.h" class LogHandler; -class Table; +class BaseTable; class Db; class MvccTrxKit; class MvccTrx; @@ -115,12 +115,12 @@ class MvccTrxLogHandler final /** * @brief 记录插入一条记录的日志 */ - RC insert_record(int32_t trx_id, Table *table, const RID &rid); + RC insert_record(int32_t trx_id, BaseTable *table, const RID &rid); /** * @brief 记录删除一条记录的日志 */ - RC delete_record(int32_t trx_id, Table *table, const RID &rid); + RC delete_record(int32_t trx_id, BaseTable *table, const RID &rid); /** * @brief 记录提交事务的日志 @@ -161,4 +161,4 @@ class MvccTrxLogReplayer final : public LogReplayer ///< 事务ID到事务的映射。在重做结束后，如果还有未提交的事务，需要回滚。 unordered_map trx_map_; -}; \ No newline at end of file +}; diff --git a/src/observer/storage/trx/trx.cpp b/src/observer/storage/trx/trx.cpp index 14ecfa78..2c35eff5 100644 --- a/src/observer/storage/trx/trx.cpp +++ b/src/observer/storage/trx/trx.cpp @@ -42,6 +42,6 @@ TrxKit *TrxKit::create(const char *name) delete trx_kit; trx_kit = nullptr; } - + return trx_kit; } diff --git a/src/observer/storage/trx/trx.h b/src/observer/storage/trx/trx.h index 5ff9af01..f23a0448 100644 --- a/src/observer/storage/trx/trx.h +++ b/src/observer/storage/trx/trx.h @@ -57,23 +57,29 @@ class Operation }; public: - Operation(Type type, Table *table, const RID &rid) - : type_(type), table_(table), page_num_(rid.page_num), slot_num_(rid.slot_num) + Operation(Type type, BaseTable *table, const RID &rid) : type_(type), table_(table), rid_(rid) {} + Operation(Type type, BaseTable *table, const RID &rid, Record &old_record, Record &updated_record) + : type_(type), table_(table), rid_(rid), old_record_(old_record), updated_record_(updated_record) {} - Type type() const { return type_; } - int32_t table_id() const { return table_->table_id(); } - Table *table() const { return table_; } - PageNum page_num() const { return page_num_; } - SlotNum slot_num() const { return slot_num_; } + Type type() const { return type_; } + int32_t table_id() const { return table_->table_id(); } + BaseTable *table() const { return table_; } + RID rid() const { return rid_; } + PageNum page_num() const { return rid_.page_num; } + SlotNum slot_num() const { return rid_.slot_num; } + const Record &old_record() const { return old_record_; } + const Record &updated_record() const { return updated_record_; } private: ///< 操作的哪张表。这里直接使用表其实并不准确，因为表中的索引也可能有日志 Type type_; - Table *table_ = nullptr; - PageNum page_num_; // TODO use RID instead of page num and slot num - SlotNum slot_num_; + BaseTable *table_ = nullptr; + RID rid_; + // update + Record old_record_{}; + Record updated_record_{}; }; class OperationHasher @@ -143,9 +149,10 @@ class Trx Trx() = default; virtual ~Trx() = default; - virtual RC insert_record(Table *table, Record &record) = 0; - virtual RC delete_record(Table *table, Record &record) = 0; - virtual RC visit_record(Table *table, Record &record, ReadWriteMode mode) = 0; + virtual RC insert_record(BaseTable *table, Record &record) = 0; + virtual RC delete_record(BaseTable *table, Record &record) = 0; + virtual RC update_record(BaseTable *table, Record &old_record, Record &new_record) = 0; + virtual RC visit_record(BaseTable *table, Record &record, ReadWriteMode mode) = 0; virtual RC start_if_need() = 0; virtual RC commit() = 0; diff --git a/src/observer/storage/trx/vacuous_trx.cpp b/src/observer/storage/trx/vacuous_trx.cpp index 1b67a46a..44a8dae5 100644 --- a/src/observer/storage/trx/vacuous_trx.cpp +++ b/src/observer/storage/trx/vacuous_trx.cpp @@ -32,11 +32,16 @@ LogReplayer *VacuousTrxKit::create_log_replayer(Db &, LogHandler &) { return new //////////////////////////////////////////////////////////////////////////////// -RC VacuousTrx::insert_record(Table *table, Record &record) { return table->insert_record(record); } +RC VacuousTrx::insert_record(BaseTable *table, Record &record) { return table->insert_record(record); } -RC VacuousTrx::delete_record(Table *table, Record &record) { return table->delete_record(record); } +RC VacuousTrx::delete_record(BaseTable *table, Record &record) { return table->delete_record(record); } -RC VacuousTrx::visit_record(Table *table, Record &record, ReadWriteMode) { return RC::SUCCESS; } +RC VacuousTrx::update_record(BaseTable *table, Record &old_record, Record &new_record) +{ + return table->update_record(old_record, new_record); +} + +RC VacuousTrx::visit_record(BaseTable *table, Record &record, ReadWriteMode) { return RC::SUCCESS; } RC VacuousTrx::start_if_need() { return RC::SUCCESS; } diff --git a/src/observer/storage/trx/vacuous_trx.h b/src/observer/storage/trx/vacuous_trx.h index f558c19c..fb38fdef 100644 --- a/src/observer/storage/trx/vacuous_trx.h +++ b/src/observer/storage/trx/vacuous_trx.h @@ -44,9 +44,10 @@ class VacuousTrx : public Trx VacuousTrx() = default; virtual ~VacuousTrx() = default; - RC insert_record(Table *table, Record &record) override; - RC delete_record(Table *table, Record &record) override; - RC visit_record(Table *table, Record &record, ReadWriteMode mode) override; + RC insert_record(BaseTable *table, Record &record) override; + RC delete_record(BaseTable *table, Record &record) override; + RC update_record(BaseTable *table, Record &old_record, Record &new_record) override; + RC visit_record(BaseTable *table, Record &record, ReadWriteMode mode) override; RC start_if_need() override; RC commit() override; RC rollback() override; @@ -63,4 +64,4 @@ class VacuousTrxLogReplayer : public LogReplayer virtual ~VacuousTrxLogReplayer() = default; RC replay(const LogEntry &) override { return RC::SUCCESS; } -}; \ No newline at end of file +}; diff --git a/test/case/miniob_test.py b/test/case/miniob_test.py index 752aed06..4d921fa1 100755 --- a/test/case/miniob_test.py +++ b/test/case/miniob_test.py @@ -1,22 +1,23 @@ # -*- coding: UTF-8 -*- -import os import json -import sys import logging -import subprocess -import socket +import os import select -import time import shutil +import socket +import subprocess +import sys import tempfile -from typing import List, Tuple +import time from enum import Enum +from typing import List, Tuple + try: - from argparse import ArgumentParser + from argparse import ArgumentParser except: - print("cannot load argparse module") - exit(1) + print("cannot load argparse module") + exit(1) _logger = logging.getLogger('MiniOBTest') @@ -46,1039 +47,1077 @@ 运行 basic 测试用例 python3 miniob_test.py --test-cases=basic - +python3 miniob_test.py --test-cases=primary-order-by +python3 miniob_test.py --test-cases=primary-date +python3 miniob_test.py --test-cases=primary-complex-sub-query +python3 miniob_test.py --test-cases=primary-simple-sub-query +python3 miniob_test.py --test-cases=vectorized-basic 如果要运行多个测试用例，则在 --test-cases 参数中使用 ',' 分隔写多个即可 """ + class TimeoutException(BaseException): - def __init__(self, value="Timed Out"): - self.value = value + def __init__(self, value="Timed Out"): + self.value = value + + def __str__(self): + return repr(self.value) - def __str__(self): - return repr(self.value) class Result(Enum): - true = True - false = False - timeout = 0 + true = True + false = False + timeout = 0 + class GlobalConfig: - default_encoding = "UTF-8" - debug = False - source_code_build_path_name = "build" + default_encoding = "UTF-8" + debug = False + source_code_build_path_name = "build" + def __get_build_path(work_dir: str): - return work_dir + '/' + GlobalConfig.source_code_build_path_name + return work_dir + '/' + GlobalConfig.source_code_build_path_name + class ResultWriter: - ''' - 写数据到指定文件，当前用于输出测试结果 - ''' + ''' + 写数据到指定文件，当前用于输出测试结果 + ''' - def __init__(self, file): - self.__file = file + def __init__(self, file): + self.__file = file - def __exit__(self, exc_type, exc_value, exc_tb): - self.close() - - def close(self): - if self.__file is not None: - self.__file.close() - self.__file = None + def __exit__(self, exc_type, exc_value, exc_tb): + self.close() - def write(self, arg: str): - self.__file.write(bytes(arg.upper(), GlobalConfig.default_encoding)) + def close(self): + if self.__file is not None: + self.__file.close() + self.__file = None - def write_line(self, arg: str): - self.write(str(arg).upper()) - self.write('\n') + def write(self, arg: str): + self.__file.write(bytes(arg.upper(), GlobalConfig.default_encoding)) -class MiniObServer: - ''' - 用来控制miniob的服务器程序。负责程序的启停和环境的初始化和清理工作 - ''' + def write_line(self, arg: str): + self.write(str(arg).upper()) + self.write('\n') - def __init__(self, base_dir: str, data_dir: str, config_file: str, server_port: int, server_socket: str, clean_data_dir: bool): - self.__check_base_dir(base_dir) - self.__check_data_dir(data_dir, clean_data_dir) - self.__base_dir = base_dir - self.__data_dir = data_dir +class MiniObServer: + ''' + 用来控制miniob的服务器程序。负责程序的启停和环境的初始化和清理工作 + ''' + + def __init__(self, base_dir: str, data_dir: str, config_file: str, server_port: int, server_socket: str, + clean_data_dir: bool): + self.__check_base_dir(base_dir) + self.__check_data_dir(data_dir, clean_data_dir) + + self.__base_dir = base_dir + self.__data_dir = data_dir + + if config_file == None: + config_file = self.__default_config(base_dir) + self.__check_config(config_file) + self.__config = config_file + self.__server_port = server_port + self.__server_socket = server_socket.strip() + + self.__process = None + + def __enter__(self): + return self + + def __exit__(self, exc_type, exc_value, exc_tb): + if self.__process is not None: + self.stop_server() + self.clean() + self.__process = None + + def __observer_path(self, base_dir: str): + ''' + observer程序所在路径 + ''' + return base_dir + "/bin/observer" + + def __default_config(self, base_dir: str): + return base_dir + "/etc/observer.ini" + + def __check_base_dir(self, base_dir: str): + if not (os.path.isdir(base_dir)): + raise (Exception("failed to check base directory. " + base_dir + " is not a directory")) + + observer_path = self.__observer_path(base_dir) + if not (os.path.isfile(observer_path)): + raise (Exception("observer not exists: " + observer_path)) + + def __check_data_dir(self, data_dir: str, clean_data_dir: bool): + if os.path.exists(data_dir) and clean_data_dir: + shutil.rmtree(data_dir) + + os.makedirs(data_dir, exist_ok=True) + if not (os.path.isdir(data_dir)): + raise (Exception(data_dir + " is not a directory or failed to create")) + + # results = os.listdir(data_dir) + # if len(results) != 0: + # raise(Exception(data_dir + " is not empty")) + + def __check_config(self, config_file: str): + if not (os.path.isfile(config_file)): + raise (Exception("config file does not exists: " + config_file)) + + def init_server(self): + _logger.info("miniob-server inited") + # do nothing now + + def start_server(self) -> bool: + ''' + 启动服务端程序，并使用探测端口的方式检测程序是否正常启动 + 调试模式如果可以使用调试器启动程序就好了 + ''' + + if self.__process != None: + _logger.warn("Server has already been started") + return False + + time_begin = time.time() + _logger.debug("use '%s' as observer work path", os.getcwd()) + observer_command = [self.__observer_path(self.__base_dir), '-f', self.__config] + if len(self.__server_socket) > 0: + observer_command.append('-s') + observer_command.append(self.__server_socket) + else: + observer_command.append('-p') + observer_command.append(str(self.__server_port)) + + process = subprocess.Popen(observer_command, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, + cwd=self.__data_dir) + return_code = process.poll() + if return_code != None: + _logger.error("Failed to start observer, exit with code %d", return_code) + return False + + _logger.info('start subprocess with pid=%d', process.pid) + # os.setpgid(process.pid, GlobalConfig.group_id) + + self.__process = process + time.sleep(0.2) + if not self.__wait_server_started(10): + time_span = time.time() - time_begin + _logger.error("Failed to start server in %f seconds", time_span) + return False + + time_span = time.time() - time_begin + _logger.info("miniob-server started in %f seconds", time_span) + return True - if config_file == None: - config_file = self.__default_config(base_dir) - self.__check_config(config_file) - self.__config = config_file - self.__server_port = server_port - self.__server_socket = server_socket.strip() + def stop_server(self): + if self.__process == None: + _logger.warning("Server has not been started") + return True + + self.__process.terminate() + return_code = -1 + try: + return_code = self.__process.wait(10) + if return_code is None: + self.__process.kill() + _logger.warning("Failed to stop server: %s", self.__base_dir) + return False + except Exception as ex: + self.__process.kill() + _logger.warning("wait server exit timedout: %s", self.__base_dir) + return False + + _logger.info("miniob-server exit with code %d. pid=%s", return_code, str(self.__process.pid)) + return True - self.__process = None + def clean(self): + ''' + 清理数据目录（如果没有配置调试模式） + 调试模式可能需要查看服务器程序运行的日志 + ''' + + if GlobalConfig.debug is False: + shutil.rmtree(self.__data_dir) + _logger.info("miniob-server cleaned") + + def __check_unix_socket_server(self): + with socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) as s: + errno = s.connect_ex(self.__server_socket) + if errno == 0: + return True + else: + _logger.debug("Failed to connect to server. err=%d:%s", errno, os.strerror(errno)) + return False + + def __check_tcp_socket_server(self): + with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s: + errno = s.connect_ex(('127.0.0.1', self.__server_port)) + if errno == 0: + return True + else: + _logger.debug("Failed to connect to server. err=%d:%s", errno, os.strerror(errno)) + return False + + def __wait_server_started(self, timeout_seconds: int): + deadline = time.time() + timeout_seconds + + while time.time() <= deadline: + result = False + if len(self.__server_socket) > 0: + result = self.__check_unix_socket_server() + else: + result = self.__check_tcp_socket_server() + if result: + return result + time.sleep(0.5) - def __enter__(self): - return self + return False - def __exit__(self, exc_type, exc_value, exc_tb): - if self.__process is not None: - self.stop_server() - self.clean() - self.__process = None - def __observer_path(self, base_dir: str): - ''' - observer程序所在路径 - ''' - return base_dir + "/bin/observer" - - def __default_config(self, base_dir: str): - return base_dir + "/etc/observer.ini" - - def __check_base_dir(self, base_dir: str): - if not(os.path.isdir(base_dir)): - raise(Exception("failed to check base directory. " + base_dir + " is not a directory")) - - observer_path = self.__observer_path(base_dir) - if not(os.path.isfile(observer_path)): - raise(Exception("observer not exists: " + observer_path)) - - def __check_data_dir(self, data_dir: str, clean_data_dir: bool): - if os.path.exists(data_dir) and clean_data_dir: - shutil.rmtree(data_dir) - - os.makedirs(data_dir, exist_ok=True) - if not(os.path.isdir(data_dir)): - raise(Exception(data_dir + " is not a directory or failed to create")) - - # results = os.listdir(data_dir) - # if len(results) != 0: - # raise(Exception(data_dir + " is not empty")) - - def __check_config(self, config_file: str): - if not(os.path.isfile(config_file)): - raise(Exception("config file does not exists: " + config_file)) - - def init_server(self): - _logger.info("miniob-server inited") - # do nothing now - - def start_server(self) -> bool: +class MiniObClient: ''' - 启动服务端程序，并使用探测端口的方式检测程序是否正常启动 - 调试模式如果可以使用调试器启动程序就好了 + 测试客户端。使用TCP连接，向服务器发送命令并反馈结果 ''' - if self.__process != None: - _logger.warn("Server has already been started") - return False + def __init__(self, server_port: int, server_socket: str, time_limit: int = 10): + if (server_port < 0 or server_port > 65535) and server_socket is None: + raise (Exception("Invalid server port: " + str(server_port))) - time_begin = time.time() - _logger.debug("use '%s' as observer work path", os.getcwd()) - observer_command = [self.__observer_path(self.__base_dir), '-f', self.__config] - if len(self.__server_socket) > 0: - observer_command.append('-s') - observer_command.append(self.__server_socket) - else: - observer_command.append('-p') - observer_command.append(str(self.__server_port)) - - process = subprocess.Popen(observer_command, stdout=subprocess.DEVNULL, stderr=subprocess.DEVNULL, cwd=self.__data_dir) - return_code = process.poll() - if return_code != None: - _logger.error("Failed to start observer, exit with code %d", return_code) - return False - - _logger.info('start subprocess with pid=%d', process.pid) - #os.setpgid(process.pid, GlobalConfig.group_id) - - self.__process = process - time.sleep(0.2) - if not self.__wait_server_started(10): - time_span = time.time() - time_begin - _logger.error("Failed to start server in %f seconds", time_span) - return False - - time_span = time.time() - time_begin - _logger.info("miniob-server started in %f seconds", time_span) - return True + self.__server_port = server_port + self.__server_socket = server_socket.strip() + self.__socket = None + self.__buffer_size = 8192 - def stop_server(self): - if self.__process == None: - _logger.warning("Server has not been started") - return True + sock = None + if len(self.__server_socket) > 0: + sock = self.__init_unix_socket(self.__server_socket) + else: + sock = self.__init_tcp_socket(self.__server_port) + + self.__socket = sock + if sock != None: + self.__socket.setblocking(False) + # self.__socket.settimeout(time_limit) # do not work + + self.__time_limit = time_limit + self.__poller = select.poll() + self.__poller.register(self.__socket, select.POLLIN | select.POLLPRI | select.POLLHUP | select.POLLERR) + + def __init_tcp_socket(self, server_port: int): + s = socket.socket(socket.AF_INET, socket.SOCK_STREAM) + errno = s.connect_ex(('127.0.0.1', server_port)) + if errno != 0: + _logger.error("Failed to connect to server with port %d. errno=%d:%s", + server_port, errno, os.strerror(errno)) + s = None + return s + + def __init_unix_socket(self, server_socket: str): + sock = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) + errno = sock.connect_ex(server_socket) + if errno != 0: + _logger.error("Failed to connect to server with address '%s'. errno=%d:%s", + server_socket, errno, os.strerror(errno)) + sock = None + return sock + + def is_valid(self): + return self.__socket is not None + + def __recv_response(self): + result = '' + + while True: + events = self.__poller.poll(self.__time_limit * 1000) + if len(events) == 0: + raise Exception('Poll timeout after %d second(s)' % self.__time_limit) + + (_, event) = events[0] + if event & (select.POLLHUP | select.POLLERR): + msg = "Failed to receive from server. poll return POLLHUP(%s) or POLLERR(%s)" % ( + str(event & select.POLLHUP), str(event & select.POLLERR)) + _logger.info(msg) + raise Exception(msg) + + data = self.__socket.recv(self.__buffer_size) + if len(data) > 0: + result_tmp = data.decode(encoding=GlobalConfig.default_encoding) + _logger.debug("receive from server[size=%d]: '%s'", len(data), result_tmp) + if data[len(data) - 1] == 0: + result += result_tmp[0:-2] + return result.strip() + '\n' + else: + result += result_tmp # TODO 返回数据量比较大的时候，python可能会hang住 + # 可以考虑返回列表 + else: + _logger.info("receive from server error. result len=%d", len(data)) + raise Exception("receive return error. the connection may be closed") + + def run_sql(self, sql: str) -> Tuple[bool, str]: + try: + data = str.encode(sql, GlobalConfig.default_encoding) + self.__socket.sendall(data) + self.__socket.sendall(b'\0') + _logger.debug("send command to server(size=%d) '%s'", len(data) + 1, sql) + result = self.__recv_response() + _logger.debug("receive result from server '%s'", result) + return True, result + except Exception as ex: + _logger.error("Failed to send message to server: '%s'", str(ex)) + return False, None + + def close(self): + if self.__socket is not None: + self.__socket.close() + self.__socket = None - self.__process.terminate() - return_code = -1 - try: - return_code = self.__process.wait(10) - if return_code is None: - self.__process.kill() - _logger.warning("Failed to stop server: %s", self.__base_dir) - return False - except Exception as ex: - self.__process.kill() - _logger.warning("wait server exit timedout: %s", self.__base_dir) - return False - _logger.info("miniob-server exit with code %d. pid=%s", return_code, str(self.__process.pid)) - return True +class CommandRunner: + __default_client_name = "default" + __command_prefix = "--" + __comment_prefix = "#" + + def __init__(self, result_writer: ResultWriter, server_port: int, unix_socket: str): + self.__result_writer = result_writer + self.__clients = {} + + # create default client + default_client = MiniObClient(server_port, unix_socket) + if not (default_client.is_valid()): + self.__is_valid = False + else: + self.__is_valid = True + self.__clients[self.__default_client_name] = default_client - def clean(self): - ''' - 清理数据目录（如果没有配置调试模式） - 调试模式可能需要查看服务器程序运行的日志 - ''' + self.__current_client = default_client + self.__server_port = server_port + self.__unix_socket = unix_socket + + def is_valid(self): + return self.__is_valid + + def __enter__(self): + return self + + def __exit__(self, exc_type, exc_value, exc_tb): + self.close() - if GlobalConfig.debug is False: - shutil.rmtree(self.__data_dir) - _logger.info("miniob-server cleaned") + def close(self): + for client in self.__clients.values(): + client.close() + self.__clients.clear() + self.__current_client = None - def __check_unix_socket_server(self): - with socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) as s: - errno = s.connect_ex(self.__server_socket) - if errno == 0: + def run_connection(self, name: str): + ''' + 切换当前连接 + ''' + + client = self.__clients[name] + if client == None: + _logger.error("No such client named %s", name) + return False + + self.__current_client = client return True - else: - _logger.debug("Failed to connect to server. err=%d:%s", errno, os.strerror(errno)) - return False - def __check_tcp_socket_server(self): - with socket.socket(socket.AF_INET, socket.SOCK_STREAM) as s: - errno = s.connect_ex(('127.0.0.1', self.__server_port)) - if errno == 0: + def run_connect(self, name: str): + ''' + 创建一个连接。每个连接有一个名字，可以使用使用connection name来切换当前的连接 + ''' + name = name.strip() + if len(name) == 0: + _logger.error("Found empty client name") + return False + + client = self.__clients[name] + if client != None: + _logger.error("Client with name %s already exists", name) + return False + + client = MiniObClient(self.__server_port, self.__unix_socket) + if not (client.is_valid()): + _logger.error("Failed to create client with name: %s", name) + return False + + self.__clients[name] = client return True - else: - _logger.debug("Failed to connect to server. err=%d:%s", errno, os.strerror(errno)) - return False - def __wait_server_started(self, timeout_seconds: int): - deadline = time.time() + timeout_seconds + def run_echo(self, arg: str): + ''' + echo 命令。参数可以是#开头的注释，这里不关心 + ''' - while time.time() <= deadline: - result = False - if len(self.__server_socket) > 0: - result = self.__check_unix_socket_server() - else: - result = self.__check_tcp_socket_server() - if result: - return result - time.sleep(0.5) + self.__result_writer.write_line(arg) + return True - return False + def run_sql(self, sql): + self.__result_writer.write_line(sql) + result, data = self.__current_client.run_sql(sql) + if result is False: + return False + self.__result_writer.write(data) + return True -class MiniObClient: - ''' - 测试客户端。使用TCP连接，向服务器发送命令并反馈结果 - ''' - - def __init__(self, server_port: int, server_socket: str, time_limit:int = 10): - if (server_port < 0 or server_port > 65535) and server_socket is None: - raise(Exception("Invalid server port: " + str(server_port))) - - self.__server_port = server_port - self.__server_socket = server_socket.strip() - self.__socket = None - self.__buffer_size = 8192 - - sock = None - if len(self.__server_socket) > 0: - sock = self.__init_unix_socket(self.__server_socket) - else: - sock = self.__init_tcp_socket(self.__server_port) - - self.__socket = sock - if sock != None: - self.__socket.setblocking(False) - #self.__socket.settimeout(time_limit) # do not work - - self.__time_limit = time_limit - self.__poller = select.poll() - self.__poller.register(self.__socket, select.POLLIN | select.POLLPRI | select.POLLHUP | select.POLLERR) - - def __init_tcp_socket(self, server_port:int): - s = socket.socket(socket.AF_INET, socket.SOCK_STREAM) - errno = s.connect_ex(('127.0.0.1', server_port)) - if errno != 0: - _logger.error("Failed to connect to server with port %d. errno=%d:%s", - server_port, errno, os.strerror(errno)) - s = None - return s - - def __init_unix_socket(self, server_socket: str): - sock = socket.socket(socket.AF_UNIX, socket.SOCK_STREAM) - errno = sock.connect_ex(server_socket) - if errno != 0: - _logger.error("Failed to connect to server with address '%s'. errno=%d:%s", - server_socket, errno, os.strerror(errno)) - sock = None - return sock - - def is_valid(self): - return self.__socket is not None - - def __recv_response(self): - result = '' - - while True: - events = self.__poller.poll(self.__time_limit * 1000) - if len(events) == 0: - raise Exception('Poll timeout after %d second(s)' % self.__time_limit) - - (_, event) = events[0] - if event & (select.POLLHUP | select.POLLERR): - msg = "Failed to receive from server. poll return POLLHUP(%s) or POLLERR(%s)" % ( str(event & select.POLLHUP), str(event & select.POLLERR)) - _logger.info(msg) - raise Exception(msg) - - data = self.__socket.recv(self.__buffer_size) - if len(data) > 0: - result_tmp = data.decode(encoding= GlobalConfig.default_encoding) - _logger.debug("receive from server[size=%d]: '%s'", len(data), result_tmp) - if data[len(data) - 1] == 0: - result += result_tmp[0:-2] - return result.strip() + '\n' + def run_sort(self, sql): + self.__result_writer.write_line(sql) + result, data = self.__current_client.run_sql(sql) + if result is False: + return False + data_l = data.strip().split('\n') + data_l.sort() + data = '\n'.join(data_l) + '\n' + self.__result_writer.write(data) + return result + + def run_command(self, command_line: str): + ''' + 执行一条命令。命令的参数使用空格分开, 第一个字符串是命令类型 + ''' + command_line = command_line[len(self.__command_prefix):] + command_line = command_line.lstrip() + args = command_line.split(' ', 1) + command = args[0] + + command_arg = '' + if len(args) > 1: + command_arg = args[1] + + result = True + if 'echo' == command: + result = self.run_echo(command_arg) + elif 'connect' == command: + result = self.run_connect(command_arg) + elif 'connection' == command: + result = self.run_connection(command_arg) + elif 'sort' == command: + result = self.run_sort(command_arg) else: - result += result_tmp # TODO 返回数据量比较大的时候，python可能会hang住 - # 可以考虑返回列表 - else: - _logger.info("receive from server error. result len=%d", len(data)) - raise Exception("receive return error. the connection may be closed") - - - def run_sql(self, sql: str) -> Tuple[bool, str]: - try: - data = str.encode(sql, GlobalConfig.default_encoding) - self.__socket.sendall(data) - self.__socket.sendall(b'\0') - _logger.debug("send command to server(size=%d) '%s'", len(data) + 1, sql) - result = self.__recv_response() - _logger.debug("receive result from server '%s'", result) - return True, result - except Exception as ex: - _logger.error("Failed to send message to server: '%s'", str(ex)) - return False, None + _logger.error("No such command %s", command) + result = False - def close(self): - if self.__socket is not None: - self.__socket.close() - self.__socket = None + return result -class CommandRunner: - __default_client_name = "default" - __command_prefix = "--" - __comment_prefix = "#" - - def __init__(self, result_writer: ResultWriter, server_port: int, unix_socket: str): - self.__result_writer = result_writer - self.__clients = {} - - # create default client - default_client = MiniObClient(server_port, unix_socket) - if not( default_client.is_valid()): - self.__is_valid = False - else: - self.__is_valid = True - self.__clients[self.__default_client_name] = default_client + def run_anything(self, argline: str): + argline = argline.strip() + if len(argline) == 0: + self.__result_writer.write_line('') # 读取到一个空行，也写入一个空行 + return True - self.__current_client = default_client - self.__server_port = server_port - self.__unix_socket = unix_socket + if argline.startswith(self.__comment_prefix): + return True - def is_valid(self): - return self.__is_valid + if argline.startswith(self.__command_prefix): + return self.run_command(argline) - def __enter__(self): - return self - - def __exit__(self, exc_type, exc_value, exc_tb): - self.close() + return self.run_sql(argline) - def close(self): - for client in self.__clients.values(): - client.close() - self.__clients.clear() - self.__current_client = None - def run_connection(self, name: str): +class TestCase: ''' - 切换当前连接 + 表示一个测试用例 + 测试用例有一个名字和内容 ''' - client = self.__clients[name] - if client == None: - _logger.error("No such client named %s", name) - return False + def __init__(self): + self.__name = '' + self.__lines = [] - self.__current_client = client - return True + def init_with_file(self, name, filename): + self.__name = name + with open(filename, mode='r') as f: + self.__lines = f.readlines() + return True - def run_connect(self, name: str): - ''' - 创建一个连接。每个连接有一个名字，可以使用使用connection name来切换当前的连接 - ''' - name = name.strip() - if len(name) == 0: - _logger.error("Found empty client name") - return False - - client = self.__clients[name] - if client != None: - _logger.error("Client with name %s already exists", name) - return False - - client = MiniObClient(self.__server_port, self.__unix_socket) - if not(client.is_valid()): - _logger.error("Failed to create client with name: %s", name) - return False - - self.__clients[name] = client - return True + def init_with_content(self, name, lines): + self.__name = name + self.__lines = lines + return True - def run_echo(self, arg: str): - ''' - echo 命令。参数可以是#开头的注释，这里不关心 - ''' + def command_lines(self): + return self.__lines - self.__result_writer.write_line(arg) - return True + def get_name(self): + return self.__name - def run_sql(self, sql): - self.__result_writer.write_line(sql) - result, data = self.__current_client.run_sql(sql) - if result is False: - return False - self.__result_writer.write(data) - return True + def result_file(self, base_dir): + subdir = '' + return base_dir + "/" + subdir + "/" + self.__name + ".result" - def run_sort(self, sql): - self.__result_writer.write_line(sql) - result, data = self.__current_client.run_sql(sql) - if result is False: - return False - data_l = data.strip().split('\n') - data_l.sort() - data = '\n'.join(data_l) + '\n' - self.__result_writer.write(data) - return result - - def run_command(self, command_line: str): + def tmp_result_file(self, base_dir): + result_file = self.result_file(base_dir) + return result_file + '.tmp' + + +class TestCaseLister: ''' - 执行一条命令。命令的参数使用空格分开, 第一个字符串是命令类型 + 列出指定目录或者指定名称的测试用例 ''' - command_line = command_line[len(self.__command_prefix) : ] - command_line = command_line.lstrip() - args = command_line.split(' ', 1) - command = args[0] - - command_arg = '' - if len(args) > 1: - command_arg = args[1] - result = True - if 'echo' == command: - result = self.run_echo(command_arg) - elif 'connect' == command: - result = self.run_connect(command_arg) - elif 'connection' == command: - result = self.run_connection(command_arg) - elif 'sort' == command: - result = self.run_sort(command_arg) - else: - _logger.error("No such command %s", command) - result = False - - return result - - def run_anything(self, argline: str): - argline = argline.strip() - if len(argline) == 0: - self.__result_writer.write_line('') # 读取到一个空行，也写入一个空行 - return True - - if argline.startswith(self.__comment_prefix): - return True - - if argline.startswith(self.__command_prefix): - return self.run_command(argline) - - return self.run_sql(argline) + def __init__(self, suffix=None): + if suffix != None: + self.__suffix = suffix + else: + self.__suffix = ".test" + + def list_directory(self, base_dir: str) -> List[TestCase]: + test_case_files = [] + + is_dir = os.path.isdir(base_dir) + if False == is_dir: + raise (Exception("Failed to list directory while getting test cases. " + base_dir + " is not a directory")) + + files = os.listdir(base_dir) + for filename in files: + _logger.debug("find file %s", filename) + if filename.startswith('.'): + continue + + full_path = base_dir + "/" + filename + is_file = os.path.isfile(full_path) + if False == is_file: + continue + if filename.endswith(self.__suffix): + test_case_files.append(filename) + + test_cases = [] + for test_case_file in test_case_files: + full_path = base_dir + "/" + test_case_file + test_case_name = test_case_file[0: -len(self.__suffix)] + test_case = TestCase() + test_case.init_with_file(test_case_name, full_path) + test_cases.append(test_case) + _logger.debug("got a test case file %s", str(test_case_file)) + + return test_cases + + def list_all(self, base_dir, test_names) -> List[TestCase]: + is_dir = os.path.isdir(base_dir) + if False == is_dir: + raise ("Failed to list all test cases. " + base_dir + " is not a directory") + + test_cases = [] + for test_name in test_names: + full_path = base_dir + "/" + test_name + self.__suffix + if not (os.path.isfile(full_path)): + raise (Exception(full_path + " is not a file")) + + test_case = TestCase() + test_case.init_with_file(test_name, full_path) + test_cases.append(test_case) + _logger.debug("got a test case %s", test_case) + + return test_cases -class TestCase: - ''' - 表示一个测试用例 - 测试用例有一个名字和内容 - ''' - - def __init__(self): - self.__name = '' - self.__lines = [] - - def init_with_file(self, name, filename): - self.__name = name - with open(filename, mode='r') as f: - self.__lines = f.readlines() - return True - def init_with_content(self, name, lines): - self.__name = name - self.__lines = lines - return True +class EvalResult: + def __init__(self): + self.__message = [] - def command_lines(self): - return self.__lines + def clear_message(self): + self.__message = [] - def get_name(self): - return self.__name + def append_message(self, message): + self.__message.append(message) - def result_file(self, base_dir): - subdir = '' - return base_dir + "/" + subdir + "/" + self.__name + ".result" + def get_message(self): + return "\n".join(self.__message) - def tmp_result_file(self, base_dir): - result_file = self.result_file(base_dir) - return result_file + '.tmp' + def to_json_string(self): + json_dict = {} + json_dict['message'] = self.get_message() -class TestCaseLister: - ''' - 列出指定目录或者指定名称的测试用例 - ''' + json_encoder = json.encoder.JSONEncoder() + json_encoder.item_separator = ',' + json_encoder.key_separator = ':' + return json_encoder.encode(json_dict) - def __init__(self, suffix = None): - if suffix != None: - self.__suffix = suffix - else: - self.__suffix = ".test" - - def list_directory(self, base_dir : str) -> List[TestCase]: - test_case_files = [] - - is_dir = os.path.isdir(base_dir) - if False == is_dir: - raise(Exception("Failed to list directory while getting test cases. " + base_dir + " is not a directory")) - - files = os.listdir(base_dir) - for filename in files: - _logger.debug("find file %s", filename) - if filename.startswith('.'): - continue - - full_path = base_dir + "/" + filename - is_file = os.path.isfile(full_path) - if False == is_file: - continue - if filename.endswith(self.__suffix): - test_case_files.append(filename) - - test_cases = [] - for test_case_file in test_case_files: - full_path = base_dir + "/" + test_case_file - test_case_name = test_case_file[0 : -len(self.__suffix)] - test_case = TestCase() - test_case.init_with_file(test_case_name, full_path) - test_cases.append(test_case) - _logger.debug("got a test case file %s", str(test_case_file)) - - return test_cases - - def list_all(self, base_dir, test_names) -> List[TestCase]: - is_dir = os.path.isdir(base_dir) - if False == is_dir: - raise("Failed to list all test cases. " + base_dir + " is not a directory") - - test_cases = [] - for test_name in test_names: - full_path = base_dir + "/" + test_name + self.__suffix - if not(os.path.isfile(full_path)): - raise(Exception(full_path + " is not a file")) - - test_case = TestCase() - test_case.init_with_file(test_name, full_path) - test_cases.append(test_case) - _logger.debug("got a test case %s", test_case) - - return test_cases -class EvalResult: - def __init__(self): - self.__message = [] - - def clear_message(self): - self.__message = [] - - def append_message(self, message): - self.__message.append(message) - - def get_message(self): - return "\n".join(self.__message) - - def to_json_string(self): - json_dict = {} - json_dict['message'] = self.get_message() - - json_encoder = json.encoder.JSONEncoder() - json_encoder.item_separator = ',' - json_encoder.key_separator = ':' - return json_encoder.encode(json_dict) - class TestSuite: - def __init__(self): - self.__report_only = False # 本次测试为了获取测试结果，不是为了校验结果 - self.__test_case_base_dir = "./test" - self.__test_result_base_dir = "./result" - self.__test_result_tmp_dir = "./result/tmp" # 生成的结果存放的临时目录 - self.__db_server_base_dir = None - self.__db_data_dir = None - self.__db_config = None - self.__server_port = 0 - self.__use_unix_socket = False # 如果指定unix socket，那么就不再使用TCP连接 - self.__need_start_server = True - self.__test_names = None # 如果指定测试哪些Case，就不再遍历所有的cases - self.__miniob_server = None - - def set_test_names(self, tests): - self.__test_names = tests - - def set_test_case_base_dir(self, test_case_base_dir): - self.__test_case_base_dir = test_case_base_dir - - def set_test_result_base_dir(self, test_result_base_dir): - self.__test_result_base_dir = test_result_base_dir - - def set_test_result_tmp_dir(self, test_result_tmp_dir: str): - self.__test_result_tmp_dir = test_result_tmp_dir - os.makedirs(test_result_tmp_dir, exist_ok=True) - if not(os.path.isdir(test_result_tmp_dir)): - raise(Exception("Failed to set test result temp directory. " + test_result_tmp_dir + " is not a directory or failed to create")) - - def set_db_server_base_dir(self, db_server_base_dir): - self.__db_server_base_dir = db_server_base_dir - - def set_db_data_dir(self, db_data_dir): - self.__db_data_dir = db_data_dir - - def set_db_config(self, db_config): - self.__db_config = db_config - - def set_server_port(self, server_port): - self.__server_port = server_port - - def set_use_unix_socket(self, use_unix_socket: bool): - self.__use_unix_socket = use_unix_socket - - def donot_need_start_server(self): - self.__need_start_server = False - - def set_report_only(self, report_only): - self.__report_only = report_only - - def __compare_files(self, file1, file2): - with open(file1, 'r') as f1, open(file2, 'r') as f2: - lines1 = f1.readlines() - lines2 = f2.readlines() - if len(lines1) != len(lines2): - return False - - line_num = len(lines1) - for i in range(line_num): - if lines1[i].upper() != lines2[i].upper(): - _logger.info('file1=%s, file2=%s, line1=%s, line2=%s', file1, file2, lines1[i], lines2[i]) - return False - return True - - def run_case(self, test_case, timeout=20) -> Result: - # eventlet.monkey_patch() - #@timeout_decorator.timeout(timeout) - #def decorator(): - try: - #with eventlet.Timeout(timeout): - ret = self.__run_case(test_case) - if ret: - return Result.true - else: - return Result.false - except TimeoutException as ex: - return Result.timeout - - def __run_case(self, test_case: TestCase) -> int: - result_tmp_file_name = test_case.tmp_result_file(self.__test_result_tmp_dir) - - unix_socket = '' - if self.__use_unix_socket: - unix_socket = self.__get_unix_socket_address() - - with open(result_tmp_file_name, mode='wb') as result_file: - result_writer = ResultWriter(result_file) - - with CommandRunner(result_writer, self.__server_port, unix_socket) as command_runner: - if command_runner.is_valid() == False: - return False - - for command_line in test_case.command_lines(): - result = command_runner.run_anything(command_line) - if result is False: - _logger.error("Failed to run command %s in case %s", command_line, test_case.get_name()) + def __init__(self): + self.__report_only = False # 本次测试为了获取测试结果，不是为了校验结果 + self.__test_case_base_dir = "./test" + self.__test_result_base_dir = "./result" + self.__test_result_tmp_dir = "./result/tmp" # 生成的结果存放的临时目录 + self.__db_server_base_dir = None + self.__db_data_dir = None + self.__db_config = None + self.__server_port = 0 + self.__use_unix_socket = False # 如果指定unix socket，那么就不再使用TCP连接 + self.__need_start_server = True + self.__test_names = None # 如果指定测试哪些Case，就不再遍历所有的cases + self.__miniob_server = None + + def set_test_names(self, tests): + self.__test_names = tests + + def set_test_case_base_dir(self, test_case_base_dir): + self.__test_case_base_dir = test_case_base_dir + + def set_test_result_base_dir(self, test_result_base_dir): + self.__test_result_base_dir = test_result_base_dir + + def set_test_result_tmp_dir(self, test_result_tmp_dir: str): + self.__test_result_tmp_dir = test_result_tmp_dir + os.makedirs(test_result_tmp_dir, exist_ok=True) + if not (os.path.isdir(test_result_tmp_dir)): + raise (Exception( + "Failed to set test result temp directory. " + test_result_tmp_dir + " is not a directory or failed to create")) + + def set_db_server_base_dir(self, db_server_base_dir): + self.__db_server_base_dir = db_server_base_dir + + def set_db_data_dir(self, db_data_dir): + self.__db_data_dir = db_data_dir + + def set_db_config(self, db_config): + self.__db_config = db_config + + def set_server_port(self, server_port): + self.__server_port = server_port + + def set_use_unix_socket(self, use_unix_socket: bool): + self.__use_unix_socket = use_unix_socket + + def donot_need_start_server(self): + self.__need_start_server = False + + def set_report_only(self, report_only): + self.__report_only = report_only + + def __compare_files(self, file1, file2): + with open(file1, 'r') as f1, open(file2, 'r') as f2: + lines1 = f1.readlines() + lines2 = f2.readlines() + + max_lines = max(len(lines1), len(lines2)) + match = True + + for i in range(max_lines): + line1 = lines1[i].strip() if i < len(lines1) else '' + line2 = lines2[i].strip() if i < len(lines2) else '' + + if line1.upper() != line2.upper(): # 使用 strip 去除行首尾空白 + _logger.info('Files are different at line %d:\n file1: %s\n file2: %s', i + 1, line1, line2) + match = False + break + + if match: + _logger.info('Files are identical: file1=%s, file2=%s', file1, file2) + else: + _logger.info('Files are different: file1=%s, file2=%s', file1, file2) + + return match + + def run_case(self, test_case, timeout=20) -> Result: + # eventlet.monkey_patch() + # @timeout_decorator.timeout(timeout) + # def decorator(): + try: + # with eventlet.Timeout(timeout): + ret = self.__run_case(test_case) + if ret: + return Result.true + else: + return Result.false + except TimeoutException as ex: + return Result.timeout + + def __run_case(self, test_case: TestCase) -> int: + result_tmp_file_name = test_case.tmp_result_file(self.__test_result_tmp_dir) + + unix_socket = '' + if self.__use_unix_socket: + unix_socket = self.__get_unix_socket_address() + + with open(result_tmp_file_name, mode='wb') as result_file: + result_writer = ResultWriter(result_file) + + with CommandRunner(result_writer, self.__server_port, unix_socket) as command_runner: + if command_runner.is_valid() == False: + return False + + for command_line in test_case.command_lines(): + result = command_runner.run_anything(command_line) + if result is False: + _logger.error("Failed to run command %s in case %s", command_line, test_case.get_name()) + return result + + result_file_name = test_case.result_file(self.__test_result_base_dir) + if self.__report_only: + os.rename(result_tmp_file_name, result_file_name) + return True + else: + result = self.__compare_files(result_tmp_file_name, result_file_name) + if not GlobalConfig.debug: + # os.remove(result_tmp_file_name) + pass return result - result_file_name = test_case.result_file(self.__test_result_base_dir) - if self.__report_only: - os.rename(result_tmp_file_name, result_file_name) - return True - else: - result = self.__compare_files(result_tmp_file_name, result_file_name) - if not GlobalConfig.debug: - #os.remove(result_tmp_file_name) - pass - return result - - def __get_unix_socket_address(self): - return self.__db_data_dir + '/miniob.sock' - - def __get_all_test_cases(self) -> List[TestCase]: - test_case_lister = TestCaseLister() - test_cases = test_case_lister.list_directory(self.__test_case_base_dir) - - if not self.__test_names: # 没有指定测试哪个case - return test_cases - - # 指定了测试case，就从中捞出来 - # 找出指定了要测试某个case，但是没有发现 - test_case_result = [] - for case_name in self.__test_names: - found = False - for test_case in test_cases: - if test_case.get_name() == case_name: - test_case_result.append(test_case) - _logger.debug("got case: " + case_name) - found = True - if found == False: - _logger.error("No such test case with name '%s'" % case_name) - return [] - - return test_case_result - - def run(self, eval_result: EvalResult): - - # 找出所有需要测试Case - test_cases = self.__get_all_test_cases() - - if not test_cases: - _logger.info("Cannot find any test cases") - return True - - _logger.info("Starting observer server") - - # 测试每个Case - success_count = 0 - failure_count = 0 - timeout_count = 0 - for test_case in test_cases: - try: - # 每个case都清理并重启一下服务端，这样可以方式某个case core之后，还能测试其它case - self.__clean_server_if_need() - - result = self.__start_server_if_need(True) - if result is False: - eval_result.append_message('Failed to start server.') - return False - - _logger.info(test_case.get_name() + " starting ...") - result = self.run_case(test_case) - - if result is Result.true: - _logger.info("Case passed: %s", test_case.get_name()) - success_count += 1 - eval_result.append_message("%s is success" % test_case.get_name()) - else: - - if result is Result.false: - _logger.info("Case failed: %s", test_case.get_name()) - failure_count += 1 - eval_result.append_message("%s is error" % test_case.get_name()) - else: - _logger.info("Case timeout: %s", test_case.get_name()) - timeout_count += 1 - eval_result.append_message("%s is timeout" % test_case.get_name()) - except Exception as ex: - _logger.error("Failed to run case %s", test_case.get_name()) + def __get_unix_socket_address(self): + return self.__db_data_dir + '/miniob.sock' + + def __get_all_test_cases(self) -> List[TestCase]: + test_case_lister = TestCaseLister() + test_cases = test_case_lister.list_directory(self.__test_case_base_dir) + + if not self.__test_names: # 没有指定测试哪个case + return test_cases + + # 指定了测试case，就从中捞出来 + # 找出指定了要测试某个case，但是没有发现 + test_case_result = [] + for case_name in self.__test_names: + found = False + for test_case in test_cases: + if test_case.get_name() == case_name: + test_case_result.append(test_case) + _logger.debug("got case: " + case_name) + found = True + if found == False: + _logger.error("No such test case with name '%s'" % case_name) + return [] + + return test_case_result + + def run(self, eval_result: EvalResult): + + # 找出所有需要测试Case + test_cases = self.__get_all_test_cases() + + if not test_cases: + _logger.info("Cannot find any test cases") + return True + + _logger.info("Starting observer server") + + # 测试每个Case + success_count = 0 + failure_count = 0 + timeout_count = 0 + for test_case in test_cases: + try: + # 每个case都清理并重启一下服务端，这样可以方式某个case core之后，还能测试其它case + self.__clean_server_if_need() + + result = self.__start_server_if_need(True) + if result is False: + eval_result.append_message('Failed to start server.') + return False + + _logger.info(test_case.get_name() + " starting ...") + result = self.run_case(test_case) + + if result is Result.true: + _logger.info("Case passed: %s", test_case.get_name()) + success_count += 1 + eval_result.append_message("%s is success" % test_case.get_name()) + else: + + if result is Result.false: + _logger.info("Case failed: %s", test_case.get_name()) + failure_count += 1 + eval_result.append_message("%s is error" % test_case.get_name()) + else: + _logger.info("Case timeout: %s", test_case.get_name()) + timeout_count += 1 + eval_result.append_message("%s is timeout" % test_case.get_name()) + except Exception as ex: + _logger.error("Failed to run case %s", test_case.get_name()) + self.__clean_server_if_need() + raise ex + + _logger.info("All done. %d passed, %d failed, %d timeout", success_count, failure_count, timeout_count) + _logger.debug(eval_result.get_message()) self.__clean_server_if_need() - raise ex + return True - _logger.info("All done. %d passed, %d failed, %d timeout", success_count, failure_count, timeout_count) - _logger.debug(eval_result.get_message()) - self.__clean_server_if_need() - return True + def __start_server_if_need(self, clean_data_dir: bool): + if self.__miniob_server is not None: + return True + + if self.__need_start_server: + unix_socket = '' + if self.__use_unix_socket: + unix_socket = self.__get_unix_socket_address() + + miniob_server = MiniObServer(self.__db_server_base_dir, self.__db_data_dir, + self.__db_config, self.__server_port, unix_socket, clean_data_dir) + miniob_server.init_server() + result = miniob_server.start_server() + if result is False: + _logger.error("Failed to start db server") + miniob_server.stop_server() + miniob_server.clean() + return False + self.__miniob_server = miniob_server - def __start_server_if_need(self, clean_data_dir: bool): - if self.__miniob_server is not None: - return True - - if self.__need_start_server: - unix_socket = '' - if self.__use_unix_socket: - unix_socket = self.__get_unix_socket_address() - - miniob_server = MiniObServer(self.__db_server_base_dir, self.__db_data_dir, - self.__db_config, self.__server_port, unix_socket, clean_data_dir) - miniob_server.init_server() - result = miniob_server.start_server() - if result is False: - _logger.error("Failed to start db server") - miniob_server.stop_server() - miniob_server.clean() - return False - self.__miniob_server = miniob_server + return True - return True + def __clean_server_if_need(self): + if self.__miniob_server is not None: + self.__miniob_server.stop_server() + # 不再清理掉中间结果。如果从解压代码开始，那么执行的中间结果不需要再清理，所有的数据都在临时目录 + # self.__miniob_server.clean() + self.__miniob_server = None - def __clean_server_if_need(self): - if self.__miniob_server is not None: - self.__miniob_server.stop_server() - # 不再清理掉中间结果。如果从解压代码开始，那么执行的中间结果不需要再清理，所有的数据都在临时目录 - # self.__miniob_server.clean() - self.__miniob_server = None def __init_options(): - options_parser = ArgumentParser() - # 是否仅仅生成结果，而不对结果做校验。一般在新生成一个case时使用 - options_parser.add_argument('--report-only', action='store_true', dest='report_only', default=False, - help='just report the result') - - # 当前miniob的代码目录 - options_parser.add_argument('--project-dir', action='store', dest='project_dir', default='') - - # 测试哪些用例。不指定就会扫描test-case-dir目录下面的所有测试用例。指定的话，就从test-case-dir目录下面按照名字找 - options_parser.add_argument('--test-cases', action='store', dest='test_cases', - help='test cases. If none, we will iterate the test case directory. Split with \',\' if more than one') - - # 测试时服务器程序的数据文件存放目录 - options_parser.add_argument('--work-dir', action='store', dest='work_dir', default='', - help='the directory of miniob database\'s data for test') - - # 服务程序端口号，客户端也使用这个端口连接服务器。目前还不具备通过配置文件解析端口配置的能力 - options_parser.add_argument('--server-port', action='store', type=int, dest='server_port', default=6789, - help='the server port. should be the same with the value in the config') - options_parser.add_argument('--not-use-unix-socket', action='store_true', dest='not_use_unix_socket', default=False, - help='If false, server-port will be ignored and will use a random address socket.') - - # 测试过程中生成的日志存放的文件。使用stdout/stderr输出到控制台 - options_parser.add_argument('--log', action='store', dest='log_file', default='stdout', - help='log file. stdout=standard output and stderr=standard error') - # 是否启动调试模式。调试模式不会清理服务器的数据目录 - options_parser.add_argument('-d', '--debug', action='store_true', dest='debug', default=False, - help='enable debug mode') - - options_parser.add_argument('--compile-make-args', action='store', dest='compile_make_args', default='', - help='compile args used by make') - options_parser.add_argument('--compile-cmake-args', action='store', dest='compile_cmake_args', default='', - help='compile args used by cmake') - # 之前已经编译过，是否需要重新编译，还是直接执行make就可以了 - options_parser.add_argument('--compile-rebuild', action='store_true', default=False, dest='compile_rebuild', - help='whether rebuild if build path exists') - - options = options_parser.parse_args(sys.argv[1:]) - - realpath = os.path.realpath(__file__) - current_path = os.path.dirname(realpath) - if not options.work_dir: - options.work_dir = tempfile.gettempdir() + '/miniob' - _logger.info('use %s as work directory', options.work_dir) - if not options.project_dir: - options.project_dir = os.path.realpath(current_path + '/../..') - _logger.info('Auto detect project dir: %s', options.project_dir) - return options + options_parser = ArgumentParser() + # 是否仅仅生成结果，而不对结果做校验。一般在新生成一个case时使用 + options_parser.add_argument('--report-only', action='store_true', dest='report_only', default=False, + help='just report the result') + + # 当前miniob的代码目录 + options_parser.add_argument('--project-dir', action='store', dest='project_dir', default='') + + # 测试哪些用例。不指定就会扫描test-case-dir目录下面的所有测试用例。指定的话，就从test-case-dir目录下面按照名字找 + options_parser.add_argument('--test-cases', action='store', dest='test_cases', + help='test cases. If none, we will iterate the test case directory. Split with \',\' if more than one') + + # 测试时服务器程序的数据文件存放目录 + options_parser.add_argument('--work-dir', action='store', dest='work_dir', default='', + help='the directory of miniob database\'s data for test') + + # 服务程序端口号，客户端也使用这个端口连接服务器。目前还不具备通过配置文件解析端口配置的能力 + options_parser.add_argument('--server-port', action='store', type=int, dest='server_port', default=6789, + help='the server port. should be the same with the value in the config') + options_parser.add_argument('--not-use-unix-socket', action='store_true', dest='not_use_unix_socket', default=False, + help='If false, server-port will be ignored and will use a random address socket.') + + # 测试过程中生成的日志存放的文件。使用stdout/stderr输出到控制台 + options_parser.add_argument('--log', action='store', dest='log_file', default='stdout', + help='log file. stdout=standard output and stderr=standard error') + # 是否启动调试模式。调试模式不会清理服务器的数据目录 + options_parser.add_argument('-d', '--debug', action='store_true', dest='debug', default=False, + help='enable debug mode') + + options_parser.add_argument('--compile-make-args', action='store', dest='compile_make_args', default='', + help='compile args used by make') + options_parser.add_argument('--compile-cmake-args', action='store', dest='compile_cmake_args', default='', + help='compile args used by cmake') + # 之前已经编译过，是否需要重新编译，还是直接执行make就可以了 + options_parser.add_argument('--compile-rebuild', action='store_true', default=False, dest='compile_rebuild', + help='whether rebuild if build path exists') + + options = options_parser.parse_args(sys.argv[1:]) + + realpath = os.path.realpath(__file__) + current_path = os.path.dirname(realpath) + if not options.work_dir: + options.work_dir = tempfile.gettempdir() + '/miniob' + _logger.info('use %s as work directory', options.work_dir) + if not options.project_dir: + options.project_dir = os.path.realpath(current_path + '/../..') + _logger.info('Auto detect project dir: %s', options.project_dir) + return options + def __init_log(options): - log_level = logging.INFO - if options.debug: - log_level = logging.DEBUG + log_level = logging.INFO + if options.debug: + log_level = logging.DEBUG + GlobalConfig.debug = True + GlobalConfig.debug = True + log_stream = None + if 'stdout' == options.log_file: + log_stream = sys.stdout + elif 'stderr' == options.log_file: + log_stream = sys.stderr + else: + log_file_dir = os.path.dirname(options.log_file) + os.makedirs(log_file_dir, exist_ok=True) - GlobalConfig.debug = True - log_stream = None - if 'stdout' == options.log_file: - log_stream = sys.stdout - elif 'stderr' == options.log_file: - log_stream = sys.stderr - else: - log_file_dir = os.path.dirname(options.log_file) - os.makedirs(log_file_dir, exist_ok=True) + log_format = "%(asctime)s - %(levelname)-5s %(name)s %(lineno)s - %(message)s" + log_date_format = "%Y-%m-%d %H:%M:%S" - log_format = "%(asctime)s - %(levelname)-5s %(name)s %(lineno)s - %(message)s" - log_date_format = "%Y-%m-%d %H:%M:%S" + if log_stream is None: + logging.basicConfig(level=log_level, filename=options.log_file, format=log_format, datefmt=log_date_format) + else: + logging.basicConfig(level=log_level, stream=log_stream, format=log_format, datefmt=log_date_format) - if log_stream is None: - logging.basicConfig(level=log_level, filename=options.log_file, format=log_format, datefmt=log_date_format) - else: - logging.basicConfig(level=log_level, stream=log_stream, format=log_format, datefmt=log_date_format) + _logger.debug('init log done') - _logger.debug('init log done') def __init_test_suite(options) -> TestSuite: - test_suite = TestSuite() - test_suite.set_test_case_base_dir(os.path.abspath(options.project_dir + '/test/case/test')) - test_suite.set_test_result_base_dir(os.path.abspath(options.project_dir + '/test/case/result')) - test_suite.set_test_result_tmp_dir(os.path.abspath(options.work_dir + '/result_output')) + test_suite = TestSuite() + test_suite.set_test_case_base_dir(os.path.abspath(options.project_dir + '/test/case/test')) + test_suite.set_test_result_base_dir(os.path.abspath(options.project_dir + '/test/case/result')) + test_suite.set_test_result_tmp_dir(os.path.abspath(options.work_dir + '/result_output')) - test_suite.set_server_port(options.server_port) - test_suite.set_use_unix_socket(not options.not_use_unix_socket) - test_suite.set_db_server_base_dir(__get_build_path(options.work_dir)) - test_suite.set_db_data_dir(options.work_dir + '/data') - test_suite.set_db_config(os.path.abspath(options.project_dir + '/etc/observer.ini')) + test_suite.set_server_port(options.server_port) + test_suite.set_use_unix_socket(not options.not_use_unix_socket) + test_suite.set_db_server_base_dir(__get_build_path(options.work_dir)) + test_suite.set_db_data_dir(options.work_dir + '/data') + test_suite.set_db_config(os.path.abspath(options.project_dir + '/etc/observer.ini')) - if options.test_cases is not None: - test_suite.set_test_names(options.test_cases.split(',')) + if options.test_cases is not None: + test_suite.set_test_names(options.test_cases.split(',')) - if options.report_only: - test_suite.set_report_only(True) + if options.report_only: + test_suite.set_report_only(True) + + return test_suite - return test_suite def __init_test_suite_with_source_code(options, eval_result): - proj_path = os.path.abspath(options.project_dir) - build_path = __get_build_path(options.work_dir) + proj_path = os.path.abspath(options.project_dir) + build_path = __get_build_path(options.work_dir) + + if not compile(proj_path, build_path, + options.compile_cmake_args, + options.compile_make_args, + options.compile_rebuild, + eval_result): + message = "Failed to compile source code" + _logger.error(message) + return None - if not compile(proj_path, build_path, - options.compile_cmake_args, - options.compile_make_args, - options.compile_rebuild, - eval_result): - message = "Failed to compile source code" - _logger.error(message) - return None + _logger.info("compile source code done") - _logger.info("compile source code done") + # 覆盖一些测试的路径 + _logger.info("some config will be override if exists") + test_suite = __init_test_suite(options) + return test_suite - # 覆盖一些测试的路径 - _logger.info("some config will be override if exists") - test_suite = __init_test_suite(options) - return test_suite def __run_shell_command(command_args): - ''' - 运行shell命令，返回命令的执行结果码和输出到控制台的信息 - 返回的控制台信息是每行一个字符串的字符串列表 - ''' - - _logger.info("running command: '%s'", ' '.join(command_args)) - - outputs = [] - command_process = subprocess.Popen(command_args, stdout=subprocess.DEVNULL, stderr=subprocess.PIPE) - while True: - line = command_process.stderr.readline() - line_str = line.decode(GlobalConfig.default_encoding) - if isinstance(line_str, str): - outputs.append(line_str.strip()) - - return_code = command_process.poll() - if return_code is not None: - return return_code, outputs + ''' + 运行shell命令，返回命令的执行结果码和输出到控制台的信息 + 返回的控制台信息是每行一个字符串的字符串列表 + ''' + + _logger.info("running command: '%s'", ' '.join(command_args)) + + outputs = [] + command_process = subprocess.Popen(command_args, stdout=subprocess.DEVNULL, stderr=subprocess.PIPE) + while True: + line = command_process.stderr.readline() + line_str = line.decode(GlobalConfig.default_encoding) + if isinstance(line_str, str): + outputs.append(line_str.strip()) + + return_code = command_process.poll() + if return_code is not None: + return return_code, outputs + def run_cmake(work_dir: str, build_path: str, cmake_args: str): - cmake_command = ["cmake", "-B", build_path, "--log-level=WARNING"] - if isinstance(cmake_args, str): - args = cmake_args.split(';') - for arg in args: - arg = arg.strip() - if len(arg) > 0: - cmake_command.append(arg) - cmake_command.append(work_dir) - - ret, outputs = __run_shell_command(cmake_command) - if ret != 0: - _logger.error("Failed to run cmake command") - for output in outputs: - _logger.error(output) - return False, outputs - return True, [] + cmake_command = ["cmake", "-B", build_path, "--log-level=WARNING"] + if isinstance(cmake_args, str): + args = cmake_args.split(';') + for arg in args: + arg = arg.strip() + if len(arg) > 0: + cmake_command.append(arg) + cmake_command.append(work_dir) + + ret, outputs = __run_shell_command(cmake_command) + if ret != 0: + _logger.error("Failed to run cmake command") + for output in outputs: + _logger.error(output) + return False, outputs + return True, [] + def compile(work_dir: str, build_dir: str, cmake_args: str, make_args: str, rebuild_all: bool, eval_result: EvalResult): - ''' - workd_dir是源代码所在目录(miniob目录) - build_dir 是编译结果的目录 - ''' - if not os.path.exists(work_dir): - _logger.error('The work_dir %s doesn\'t exist, please provide a vaild work path.', work_dir) - return False - - #now_path = os.getcwd() - build_path = build_dir - if os.path.exists(build_path) and rebuild_all: - _logger.info('build directory is not empty but will be cleaned before compile: %s', build_path) - shutil.rmtree(build_path) - - os.makedirs(build_path, exist_ok=True) - - _logger.info("start compiling ... build path=%s", build_path) - ret, outputs = run_cmake(work_dir, build_path, cmake_args) - if ret == False: - # cmake 执行失败时，清空整个Build目录，再重新执行一次cmake命令 - shutil.rmtree(build_path) + ''' + workd_dir是源代码所在目录(miniob目录) + build_dir 是编译结果的目录 + ''' + if not os.path.exists(work_dir): + _logger.error('The work_dir %s doesn\'t exist, please provide a vaild work path.', work_dir) + return False + + # now_path = os.getcwd() + build_path = build_dir + if os.path.exists(build_path) and rebuild_all: + _logger.info('build directory is not empty but will be cleaned before compile: %s', build_path) + shutil.rmtree(build_path) + os.makedirs(build_path, exist_ok=True) + + _logger.info("start compiling ... build path=%s", build_path) ret, outputs = run_cmake(work_dir, build_path, cmake_args) if ret == False: - for output in outputs: - _logger.error(output) - eval_result.append_message(output) - return False - - make_command = ["make", "--silent", "-C", build_path] - if isinstance(make_args, str): - if not make_args: - make_command.append('-j4') - else: - args = make_args.split(';') - for arg in args: - arg = arg.strip() - if len(arg) > 0: - make_command.append(arg) - - ret, outputs = __run_shell_command(make_command) - if ret != 0: - _logger.error("Compile failed") - for output in outputs: - _logger.error(output.strip()) - eval_result.append_message(output.strip()) - return False - - return True + # cmake 执行失败时，清空整个Build目录，再重新执行一次cmake命令 + shutil.rmtree(build_path) + os.makedirs(build_path, exist_ok=True) + ret, outputs = run_cmake(work_dir, build_path, cmake_args) + if ret == False: + for output in outputs: + _logger.error(output) + eval_result.append_message(output) + return False + + make_command = ["make", "--silent", "-C", build_path] + if isinstance(make_args, str): + if not make_args: + make_command.append('-j16') + else: + args = make_args.split(';') + for arg in args: + arg = arg.strip() + if len(arg) > 0: + make_command.append(arg) + + ret, outputs = __run_shell_command(make_command) + if ret != 0: + _logger.error("Compile failed") + for output in outputs: + _logger.error(output.strip()) + eval_result.append_message(output.strip()) + return False + + return True + def run(options) -> Tuple[bool, str]: - ''' - return result, reason - result: True or False - - ''' - __init_log(options) - - _logger.info("miniob test starting ...") - - # 由于miniob-test测试程序导致的失败，才认为是失败 - # 比如目录没有权限等，对miniob-test来说是成功的 - result = True - eval_result = EvalResult() - - try: - test_suite:TestSuite = __init_test_suite_with_source_code(options, eval_result) - - if test_suite != None: - result = test_suite.run(eval_result) - # result = True - except Exception as ex: - _logger.exception(ex) - result = False - #eval_result.clear_message() - eval_result.append_message(str(ex.args)) - - return result, eval_result.to_json_string() + ''' + return result, reason + result: True or False + + ''' + __init_log(options) + + _logger.info("miniob test starting ...") + + # 由于miniob-test测试程序导致的失败，才认为是失败 + # 比如目录没有权限等，对miniob-test来说是成功的 + result = True + eval_result = EvalResult() + + try: + test_suite: TestSuite = __init_test_suite_with_source_code(options, eval_result) + + if test_suite != None: + result = test_suite.run(eval_result) + # result = True + except Exception as ex: + _logger.exception(ex) + result = False + # eval_result.clear_message() + eval_result.append_message(str(ex.args)) + + return result, eval_result.to_json_string() + if __name__ == '__main__': - os.setpgrp() - options = __init_options() + os.setpgrp() + options = __init_options() - result, evaluation = run(options) + result, evaluation = run(options) - exit_code = 0 - if result is False: - exit_code = 1 - else: - _logger.info(evaluation) - exit(exit_code) + exit_code = 0 + if result is False: + exit_code = 1 + else: + _logger.info(evaluation) + exit(exit_code) diff --git a/test/case/result/basic.result b/test/case/result/basic.result index 83ef38c0..99a88aa5 100644 --- a/test/case/result/basic.result +++ b/test/case/result/basic.result @@ -128,10 +128,10 @@ calc (1+2) * (2 * (20+-(5*1))); 90 calc 20/0; 20/0 -340282346638528859811704183484516925440 +NULL calc 13.2/0.0; 13.2/0.0 -340282346638528859811704183484516925440 +NULL calc "123" + 4; "123" + 4 127 diff --git a/test/case/result/primary-aggregation-func.result b/test/case/result/primary-aggregation-func.result index c4f15205..c2c0390e 100644 --- a/test/case/result/primary-aggregation-func.result +++ b/test/case/result/primary-aggregation-func.result @@ -1,5 +1,5 @@ INITIALIZATION -CREATE TABLE aggregation_func(id int, num int, price float, addr char, birthday date); +CREATE TABLE aggregation_func(id int, num int, price float, addr char(4), birthday date); SUCCESS INSERT INTO aggregation_func VALUES (1, 18, 10.0, 'abc', '2020-01-01'); diff --git a/test/case/result/primary-complex-sub-query-plus.result b/test/case/result/primary-complex-sub-query-plus.result new file mode 100644 index 00000000..fc5fa1f4 --- /dev/null +++ b/test/case/result/primary-complex-sub-query-plus.result @@ -0,0 +1,62 @@ +INITIALIZATION +CREATE TABLE CSQ_1(ID INT, COL1 INT, FEAT1 FLOAT); +SUCCESS +CREATE TABLE CSQ_2(ID INT, COL2 INT, FEAT2 FLOAT); +SUCCESS +CREATE TABLE CSQ_3(ID INT, COL3 INT, FEAT3 FLOAT); +SUCCESS + +INSERT INTO CSQ_1 VALUES (1, 4, 11.2); +SUCCESS +INSERT INTO CSQ_1 VALUES (2, 2, 12.0); +SUCCESS +INSERT INTO CSQ_1 VALUES (3, 3, 13.5); +SUCCESS +INSERT INTO CSQ_1 VALUES (4, 1, 10.8); +SUCCESS +INSERT INTO CSQ_1 VALUES (5, 5, 15.3); +SUCCESS +INSERT INTO CSQ_1 VALUES (6, 7, 12.4); +SUCCESS +INSERT INTO CSQ_1 VALUES (7, 6, 11.9); +SUCCESS + +INSERT INTO CSQ_2 VALUES (1, 2, 13.0); +SUCCESS +INSERT INTO CSQ_2 VALUES (2, 7, 10.5); +SUCCESS +INSERT INTO CSQ_2 VALUES (3, 9, 13.8); +SUCCESS +INSERT INTO CSQ_2 VALUES (4, 5, 14.1); +SUCCESS +INSERT INTO CSQ_2 VALUES (5, 3, 12.6); +SUCCESS +INSERT INTO CSQ_2 VALUES (6, 1, 10.2); +SUCCESS +INSERT INTO CSQ_2 VALUES (7, 4, 12.3); +SUCCESS + +INSERT INTO CSQ_3 VALUES (1, 2, 11.0); +SUCCESS +INSERT INTO CSQ_3 VALUES (2, 3, 13.4); +SUCCESS +INSERT INTO CSQ_3 VALUES (3, 6, 16.5); +SUCCESS +INSERT INTO CSQ_3 VALUES (4, 4, 12.9); +SUCCESS +INSERT INTO CSQ_3 VALUES (5, 5, 14.6); +SUCCESS +INSERT INTO CSQ_3 VALUES (6, 7, 11.7); +SUCCESS +INSERT INTO CSQ_3 VALUES (7, 8, 15.0); +SUCCESS + +1. SELECT +SELECT * FROM CSQ_1 WHERE FEAT1 <> (SELECT MIN(CSQ_2.FEAT2) FROM CSQ_2 WHERE CSQ_2.FEAT2 > CSQ_1.FEAT1); +1 | 4 | 11.2 +2 | 2 | 12 +3 | 3 | 13.5 +4 | 1 | 10.8 +6 | 7 | 12.4 +7 | 6 | 11.9 +ID | COL1 | FEAT1 diff --git a/test/case/result/primary-complex-sub-query.result b/test/case/result/primary-complex-sub-query.result index d499849b..867495e2 100644 --- a/test/case/result/primary-complex-sub-query.result +++ b/test/case/result/primary-complex-sub-query.result @@ -77,6 +77,9 @@ select * from csq_1 where feat1 <> (select avg(csq_2.feat2) from csq_2 where csq 2 | 2 | 12 ID | COL1 | FEAT1 +select * from csq_1 where feat1 > (select min(csq_2.feat2) from csq_2 where csq_2.feat2 > csq_1.feat1); +ID | COL1 | FEAT1 + select * from csq_1 where col1 not in (select csq_2.col2 from csq_2 where csq_2.id in (select csq_3.id from csq_3 where csq_1.id = csq_3.id)); 1 | 4 | 11.2 2 | 2 | 12 diff --git a/test/case/result/primary-create-view.result b/test/case/result/primary-create-view.result new file mode 100644 index 00000000..e69de29b diff --git a/test/case/result/primary-drop-table.result b/test/case/result/primary-drop-table.result index 39b2331f..d4307bf9 100644 --- a/test/case/result/primary-drop-table.result +++ b/test/case/result/primary-drop-table.result @@ -1,11 +1,11 @@ 1. DROP EMPTY TABLE -CREATE TABLE Drop_table_1(id int, t_name char); +CREATE TABLE Drop_table_1(id int, t_name char(4)); SUCCESS DROP TABLE Drop_table_1; SUCCESS 2. DROP NON-EMPTY TABLE -CREATE TABLE Drop_table_2(id int, t_name char); +CREATE TABLE Drop_table_2(id int, t_name char(4)); SUCCESS INSERT INTO Drop_table_2 VALUES (1,'OB'); SUCCESS @@ -13,7 +13,7 @@ DROP TABLE Drop_table_2; SUCCESS 3. CHECK THE ACCURACY OF DROPPING TABLE -CREATE TABLE Drop_table_3(id int, t_name char); +CREATE TABLE Drop_table_3(id int, t_name char(4)); SUCCESS INSERT INTO Drop_table_3 VALUES (1,'OB'); SUCCESS @@ -28,13 +28,13 @@ SELECT * FROM Drop_table_3; FAILURE DELETE FROM Drop_table_3 WHERE id = 3; FAILURE -CREATE TABLE Drop_table_3(id int, t_name char); +CREATE TABLE Drop_table_3(id int, t_name char(4)); SUCCESS SELECT * FROM Drop_table_3; ID | T_NAME 4. DROP NON-EXISTENT TABLE -CREATE TABLE Drop_table_4(id int, t_name char); +CREATE TABLE Drop_table_4(id int, t_name char(4)); SUCCESS DROP TABLE Drop_table_4; SUCCESS @@ -44,17 +44,17 @@ DROP TABLE Drop_table_4_1; FAILURE 5. CREATE A TABLE WHICH HAS DROPPED -CREATE TABLE Drop_table_5(id int, t_name char); +CREATE TABLE Drop_table_5(id int, t_name char(4)); SUCCESS DROP TABLE Drop_table_5; SUCCESS -CREATE TABLE Drop_table_5(id int, t_name char); +CREATE TABLE Drop_table_5(id int, t_name char(4)); SUCCESS SELECT * FROM Drop_table_5; ID | T_NAME 6. DROP A TABLE WITH INDEX -CREATE TABLE Drop_table_6(id int, t_name char); +CREATE TABLE Drop_table_6(id int, t_name char(4)); SUCCESS CREATE INDEX index_id on Drop_table_6(id); SUCCESS diff --git a/test/case/result/primary-function.result b/test/case/result/primary-function.result new file mode 100644 index 00000000..7419f9bd --- /dev/null +++ b/test/case/result/primary-function.result @@ -0,0 +1,38 @@ +1. INIT DATA +CREATE TABLE FUNCTION_TABLE(ID INT, NAME CHAR(20), SCORE FLOAT, U_DATE DATE); +SUCCESS +CREATE TABLE FUNCTION_TABLE_2(ID INT, NAME CHAR(30), SCORE FLOAT, U_DATE DATE); +SUCCESS +INSERT INTO FUNCTION_TABLE VALUES (4, 'KI83C4RH0AGX9', 7.61, '2017-02-06'); +SUCCESS +INSERT INTO FUNCTION_TABLE VALUES (4, 'FXMB2DOJ4', 8.35, '2021-10-07'); +SUCCESS +INSERT INTO FUNCTION_TABLE VALUES (8, 'Q3NS46CJ', 8.66, '2023-06-19'); +SUCCESS +INSERT INTO FUNCTION_TABLE VALUES (1, 'Q3NS4', 4.90, '2023-06-19'); +SUCCESS +INSERT INTO FUNCTION_TABLE VALUES (3, 'Q3NS46CJ012311234561', 6.50, '2023-06-19'); +SUCCESS +INSERT INTO FUNCTION_TABLE VALUES (2, 'Q3CJ', 9.00, '2023-06-19'); +SUCCESS + +2. INVALID DATE +SELECT ID, DATE_FORMAT(NAME, '%Y-%M-%D') FROM FUNCTION_TABLE; +FAILURE + +3. ROUND TEST +SELECT ID, LENGTH(NAME), ROUND(SCORE) FROM FUNCTION_TABLE WHERE ID<8; +1 | 5 | 5 +2 | 4 | 9 +3 | 20 | 7 +4 | 13 | 8 +4 | 9 | 8 +ID | LENGTH(NAME) | ROUND(SCORE) +SELECT * FROM FUNCTION_TABLE; +1 | Q3NS4 | 4.9 | 2023-06-19 +2 | Q3CJ | 9 | 2023-06-19 +3 | Q3NS46CJ012311234561 | 6.5 | 2023-06-19 +4 | FXMB2DOJ4 | 8.35 | 2021-10-07 +4 | KI83C4RH0AGX9 | 7.61 | 2017-02-06 +8 | Q3NS46CJ | 8.66 | 2023-06-19 +ID | NAME | SCORE | U_DATE diff --git a/test/case/result/primary-group-by.result b/test/case/result/primary-group-by.result index cec89e77..6ba93d97 100644 --- a/test/case/result/primary-group-by.result +++ b/test/case/result/primary-group-by.result @@ -1,7 +1,7 @@ 1. CREATE TABLE -create table t_group_by (id int, score float, name char); +create table t_group_by (id int not null, score float not null, name char(1) null); SUCCESS -create table t_group_by_2 (id int, age int); +create table t_group_by_2 (id int not null, age int not null); SUCCESS 2. INSERT RECORDS @@ -19,6 +19,8 @@ insert into t_group_by values(3, 3.0, 'd'); SUCCESS insert into t_group_by values(3, 2.0, 'f'); SUCCESS +INSERT INTO T_GROUP_BY VALUES(4, 3.0, NULL); +SUCCESS insert into t_group_by_2 values(1, 10); SUCCESS @@ -40,7 +42,14 @@ select id, avg(score) from t_group_by group by id; 4 | 3 ID | AVG(SCORE) +select id, sum(score) from t_group_by group by id; +1 | 2 +3 | 12 +4 | 6 +ID | SUM(SCORE) + select name, min(id), max(score) from t_group_by group by name; +NULL | 4 | 3 A | 3 | 1 B | 1 | 2 C | 3 | 4 @@ -54,6 +63,7 @@ select id, name, avg(score) from t_group_by group by id, name; 3 | C | 3 3 | D | 3 3 | F | 2 +4 | NULL | 3 4 | C | 3 ID | NAME | AVG(SCORE) @@ -77,5 +87,12 @@ select t_group_by.id, t_group_by.name, avg(t_group_by.score), avg(t_group_by_2.a 3 | C | 3 | 23.33 3 | D | 3 | 23.33 3 | F | 2 | 23.33 +4 | NULL | 3 | 20 4 | C | 3 | 20 T_GROUP_BY.ID | T_GROUP_BY.NAME | AVG(T_GROUP_BY.SCORE) | AVG(T_GROUP_BY_2.AGE) + +6. HAVING TYPECAST +SELECT ID, AVG(SCORE) FROM T_GROUP_BY GROUP BY ID HAVING AVG(SCORE)>2; +3 | 2.4 +4 | 3 +ID | AVG(SCORE) diff --git a/test/case/result/primary-insert.result b/test/case/result/primary-insert.result index cb054382..bfec2aeb 100644 --- a/test/case/result/primary-insert.result +++ b/test/case/result/primary-insert.result @@ -1,22 +1,54 @@ INITIALIZATION -CREATE TABLE insert_table(id int, t_name char, col1 int, col2 int); +CREATE TABLE INSERT_TABLE(ID INT, T_NAME CHAR(4), COL1 INT, COL2 INT); +SUCCESS +CREATE TABLE INSERT_TABLE2(ID INT, T_NAME CHAR(6), COL1 INT, COL2 INT); SUCCESS 1. INSERT -INSERT INTO insert_table VALUES (1,'N1',1,1); +INSERT INTO INSERT_TABLE VALUES (1,'N1',1,1); +SUCCESS +INSERT INTO INSERT_TABLE VALUES (1,'N123',1,1); +SUCCESS +INSERT INTO INSERT_TABLE VALUES (2,'N123',1,1),(3,'N32',2,1); +SUCCESS + +INSERT INTO INSERT_TABLE2 VALUES (1,'N1',1,1); +SUCCESS +INSERT INTO INSERT_TABLE2 VALUES (1,'N123',1,1); SUCCESS -INSERT INTO insert_table VALUES (2,'N2',1,1),(3,'N3',2,1); +INSERT INTO INSERT_TABLE2 VALUES (2,'N1233',1,1),(3,'N3211',2,1); +SUCCESS +INSERT INTO INSERT_TABLE2 VALUES (4,'','',''); SUCCESS 2. ERROR -INSERT INTO insert_table VALUES (4,'N4',1,1),(1,1,1); +INSERT INTO INSERT_TABLE VALUES (5,'N4',1,1),(1,1,1); +FAILURE +INSERT INTO INSERT_TABLE VALUES (5,'N4',1,1),(1,1,1,1,1); FAILURE -INSERT INTO insert_table VALUES (4,'N4',1,1),(1,1,1,1); +INSERT INTO INSERT_TABLE VALUES (5,'12345',1,1); +FAILURE + +INSERT INTO INSERT_TABLE2 VALUES (5,'N412',1,1),(1,1,1); +FAILURE +INSERT INTO INSERT_TABLE2 VALUES (5,'N41',1,1),(1,1,1,1,1); +FAILURE +INSERT INTO INSERT_TABLE2 VALUES (5,'1234567',1,1); +FAILURE +INSERT INTO INSERT_TABLE2 VALUES (5,'12345678910',1,1); FAILURE 3. SELECT -SELECT * FROM insert_table; +SELECT * FROM INSERT_TABLE; +1 | N1 | 1 | 1 +1 | N123 | 1 | 1 +2 | N123 | 1 | 1 +3 | N32 | 2 | 1 +ID | T_NAME | COL1 | COL2 +SELECT * FROM INSERT_TABLE2; 1 | N1 | 1 | 1 -2 | N2 | 1 | 1 -3 | N3 | 2 | 1 +1 | N123 | 1 | 1 +2 | N1233 | 1 | 1 +3 | N3211 | 2 | 1 +4 | | 0 | 0 ID | T_NAME | COL1 | COL2 diff --git a/test/case/result/primary-null.result b/test/case/result/primary-null.result index b04d9522..b8860fa1 100644 --- a/test/case/result/primary-null.result +++ b/test/case/result/primary-null.result @@ -1,7 +1,7 @@ INITIALIZATION -CREATE TABLE null_table(id int, num int nullable, price float not null, birthday date nullable); +CREATE TABLE null_table(id int not null, num int, price float not null, birthday date null); SUCCESS -CREATE TABLE null_table2(id int, num int nullable, price float not null, birthday date nullable); +CREATE TABLE null_table2(id int not null, num int, price float not null, birthday date null); SUCCESS CREATE INDEX index_num on null_table(num); SUCCESS @@ -164,7 +164,7 @@ AVG(NUM) 15 6. AGGREGATION WITH NULL COLUMNS -CREATE TABLE null_table3(id int, num int nullable); +CREATE TABLE null_table3(id int not null, num int null); SUCCESS INSERT INTO null_table3 VALUES (1, null); SUCCESS diff --git a/test/case/result/primary-show-index.result b/test/case/result/primary-show-index.result new file mode 100644 index 00000000..1ca4c7e1 --- /dev/null +++ b/test/case/result/primary-show-index.result @@ -0,0 +1,35 @@ +INITIALIZATION +CREATE TABLE INDEX_TABLE_1(ID INT, AGE INT); +SUCCESS +INSERT INTO INDEX_TABLE_1 VALUES (1,1); +SUCCESS +INSERT INTO INDEX_TABLE_1 VALUES (2,2); +SUCCESS +INSERT INTO INDEX_TABLE_1 VALUES (3,3); +SUCCESS +INSERT INTO INDEX_TABLE_1 VALUES (1,2); +SUCCESS +INSERT INTO INDEX_TABLE_1 VALUES (1,3); +SUCCESS + +1. SHOW EMPTY INDEX +SHOW INDEX FROM INDEX_TABLE_1; +TABLE | NON_UNIQUE | KEY_NAME | SEQ_IN_INDEX | COLUMN_NAME + +2. SHOW AN INDEX +CREATE INDEX INDEX_ID_1 ON INDEX_TABLE_1(ID); +SUCCESS +SHOW INDEX FROM INDEX_TABLE_1; +TABLE | NON_UNIQUE | KEY_NAME | SEQ_IN_INDEX | COLUMN_NAME +INDEX_TABLE_1 | 1 | INDEX_ID_1 | 1 | ID + +3. SHOW INDEXES +CREATE INDEX INDEX_ID_2 ON INDEX_TABLE_1(ID); +SUCCESS +CREATE INDEX INDEX_ID_3 ON INDEX_TABLE_1(AGE); +SUCCESS +SHOW INDEX FROM INDEX_TABLE_1; +TABLE | NON_UNIQUE | KEY_NAME | SEQ_IN_INDEX | COLUMN_NAME +INDEX_TABLE_1 | 1 | INDEX_ID_1 | 1 | ID +INDEX_TABLE_1 | 1 | INDEX_ID_2 | 1 | ID +INDEX_TABLE_1 | 1 | INDEX_ID_3 | 1 | AGE diff --git a/test/case/result/primary-simple-sub-query.result b/test/case/result/primary-simple-sub-query.result index 0ffc1cfc..a683dd80 100644 --- a/test/case/result/primary-simple-sub-query.result +++ b/test/case/result/primary-simple-sub-query.result @@ -1,109 +1,117 @@ INITIALIZATION -CREATE TABLE ssq_1(id int, col1 int, feat1 float); +CREATE TABLE SSQ_1(ID INT, COL1 INT, FEAT1 FLOAT); SUCCESS -CREATE TABLE ssq_2(id int, col2 int, feat2 float); +CREATE TABLE SSQ_2(ID INT, COL2 INT, FEAT2 FLOAT); SUCCESS -CREATE TABLE ssq_3(id int, col3 int, feat3 float); +CREATE TABLE SSQ_3(ID INT, COL3 INT, FEAT3 FLOAT); SUCCESS -INSERT INTO ssq_1 VALUES (1, 4, 11.2); +INSERT INTO SSQ_1 VALUES (1, 4, 11.2); SUCCESS -INSERT INTO ssq_1 VALUES (2, 2, 12.0); +INSERT INTO SSQ_1 VALUES (2, 2, 12.0); SUCCESS -INSERT INTO ssq_1 VALUES (3, 3, 13.5); +INSERT INTO SSQ_1 VALUES (3, 3, 13.5); SUCCESS -INSERT INTO ssq_2 VALUES (1, 2, 13.0); +INSERT INTO SSQ_2 VALUES (1, 2, 13.0); SUCCESS -INSERT INTO ssq_2 VALUES (2, 7, 10.5); +INSERT INTO SSQ_2 VALUES (2, 7, 10.5); SUCCESS -INSERT INTO ssq_2 VALUES (5, 3, 12.6); +INSERT INTO SSQ_2 VALUES (5, 3, 12.6); SUCCESS 1. SELECT -select * from ssq_1 where id in (select ssq_2.id from ssq_2); +SELECT * FROM SSQ_1 WHERE ID IN (SELECT SSQ_2.ID FROM SSQ_2); 1 | 4 | 11.2 2 | 2 | 12 ID | COL1 | FEAT1 -select * from ssq_1 where col1 not in (select ssq_2.col2 from ssq_2); +SELECT * FROM SSQ_1 WHERE COL1 NOT IN (SELECT SSQ_2.COL2 FROM SSQ_2); 1 | 4 | 11.2 ID | COL1 | FEAT1 -select * from ssq_1 where col1 = (select avg(ssq_2.col2) from ssq_2); +SELECT * FROM SSQ_1 WHERE COL1 = (SELECT AVG(SSQ_2.COL2) FROM SSQ_2); 1 | 4 | 11.2 ID | COL1 | FEAT1 -select * from ssq_1 where (select avg(ssq_2.col2) from ssq_2) = col1; +SELECT * FROM SSQ_1 WHERE (SELECT AVG(SSQ_2.COL2) FROM SSQ_2) = COL1; 1 | 4 | 11.2 ID | COL1 | FEAT1 -select * from ssq_1 where feat1 >= (select min(ssq_2.feat2) from ssq_2); +SELECT * FROM SSQ_1 WHERE FEAT1 >= (SELECT MIN(SSQ_2.FEAT2) FROM SSQ_2); 1 | 4 | 11.2 2 | 2 | 12 3 | 3 | 13.5 ID | COL1 | FEAT1 -select * from ssq_1 where (select min(ssq_2.feat2) from ssq_2) <= feat1; +SELECT * FROM SSQ_1 WHERE (SELECT MIN(SSQ_2.FEAT2) FROM SSQ_2) <= FEAT1; 1 | 4 | 11.2 2 | 2 | 12 3 | 3 | 13.5 ID | COL1 | FEAT1 -select * from ssq_1 where feat1 <= (select max(ssq_2.feat2) from ssq_2); +SELECT * FROM SSQ_1 WHERE FEAT1 <= (SELECT MAX(SSQ_2.FEAT2) FROM SSQ_2); 1 | 4 | 11.2 2 | 2 | 12 ID | COL1 | FEAT1 -select * from ssq_1 where (select max(ssq_2.feat2) from ssq_2) >= feat1; +SELECT * FROM SSQ_1 WHERE (SELECT MAX(SSQ_2.FEAT2) FROM SSQ_2) >= FEAT1; 1 | 4 | 11.2 2 | 2 | 12 ID | COL1 | FEAT1 -select * from ssq_1 where feat1 > (select min(ssq_2.feat2) from ssq_2); +SELECT * FROM SSQ_1 WHERE FEAT1 > (SELECT MIN(SSQ_2.FEAT2) FROM SSQ_2); 1 | 4 | 11.2 2 | 2 | 12 3 | 3 | 13.5 ID | COL1 | FEAT1 -select * from ssq_1 where (select min(ssq_2.feat2) from ssq_2) < feat1; +SELECT * FROM SSQ_1 WHERE (SELECT MIN(SSQ_2.FEAT2) FROM SSQ_2) < FEAT1; 1 | 4 | 11.2 2 | 2 | 12 3 | 3 | 13.5 ID | COL1 | FEAT1 -select * from ssq_1 where feat1 < (select max(ssq_2.feat2) from ssq_2); +SELECT * FROM SSQ_1 WHERE FEAT1 < (SELECT MAX(SSQ_2.FEAT2) FROM SSQ_2); 1 | 4 | 11.2 2 | 2 | 12 ID | COL1 | FEAT1 -select * from ssq_1 where (select max(ssq_2.feat2) from ssq_2) > feat1; +SELECT * FROM SSQ_1 WHERE (SELECT MAX(SSQ_2.FEAT2) FROM SSQ_2) > FEAT1; 1 | 4 | 11.2 2 | 2 | 12 ID | COL1 | FEAT1 -select * from ssq_1 where feat1 <> (select avg(ssq_2.feat2) from ssq_2); +SELECT * FROM SSQ_1 WHERE FEAT1 <> (SELECT AVG(SSQ_2.FEAT2) FROM SSQ_2); 1 | 4 | 11.2 2 | 2 | 12 3 | 3 | 13.5 ID | COL1 | FEAT1 2. SELECT WITH EMPTY TABLE -select * from ssq_1 where feat1 < (select max(ssq_2.feat2) from ssq_2 where 1=0); +SELECT * FROM SSQ_1 WHERE COL1 NOT IN (2,NULL); ID | COL1 | FEAT1 -select * from ssq_1 where id in (select ssq_2.id from ssq_2 where 1=0); +SELECT * FROM SSQ_1 WHERE FEAT1 < (SELECT MAX(SSQ_2.FEAT2) FROM SSQ_2 WHERE 1=0); ID | COL1 | FEAT1 -select * from ssq_1 where id not in (select ssq_2.id from ssq_2 where 1=0); +SELECT * FROM SSQ_1 WHERE ID IN (SELECT SSQ_2.ID FROM SSQ_2 WHERE 1=0); +ID | COL1 | FEAT1 +SELECT * FROM SSQ_1 WHERE ID NOT IN (SELECT SSQ_2.ID FROM SSQ_2 WHERE 1=0); 1 | 4 | 11.2 2 | 2 | 12 3 | 3 | 13.5 ID | COL1 | FEAT1 -select * from ssq_3 where feat3 < (select max(ssq_2.feat2) from ssq_2); +SELECT * FROM SSQ_3 WHERE FEAT3 < (SELECT MAX(SSQ_2.FEAT2) FROM SSQ_2); ID | COL3 | FEAT3 -select * from ssq_3 where id in (select ssq_2.id from ssq_2); +SELECT * FROM SSQ_3 WHERE ID IN (SELECT SSQ_2.ID FROM SSQ_2); ID | COL3 | FEAT3 -select * from ssq_3 where id not in (select ssq_2.id from ssq_2); +SELECT * FROM SSQ_3 WHERE ID NOT IN (SELECT SSQ_2.ID FROM SSQ_2); ID | COL3 | FEAT3 +SELECT * FROM SSQ_1 WHERE (SELECT SSQ_2.ID FROM SSQ_2 WHERE COL2 = 52) = ID; +ID | COL1 | FEAT1 3. ERROR -select * from ssq_1 where col1 = (select ssq_2.col2 from ssq_2); +SELECT * FROM SSQ_1 WHERE COL1 = (SELECT SSQ_2.COL2 FROM SSQ_2); FAILURE -select * from ssq_1 where col1 = (select * from ssq_2); +SELECT * FROM SSQ_1 WHERE COL1 = (SELECT * FROM SSQ_2); FAILURE -select * from ssq_1 where col1 in (select * from ssq_2); +SELECT * FROM SSQ_1 WHERE COL1 IN (SELECT * FROM SSQ_2); FAILURE -select * from ssq_1 where col1 not in (select * from ssq_2); +SELECT * FROM SSQ_1 WHERE COL1 NOT IN (SELECT * FROM SSQ_2); FAILURE + +4. STREAM CLEAR AFTER ERROR +SELECT * FROM SSQ_1 WHERE (SELECT SSQ_2.ID FROM SSQ_2 WHERE COL2 = 52) = ID; +ID | COL1 | FEAT1 diff --git a/test/case/result/primary-text.result b/test/case/result/primary-text.result index 5b31e529..a8fddf11 100644 --- a/test/case/result/primary-text.result +++ b/test/case/result/primary-text.result @@ -46,5 +46,5 @@ select * from text_table; 2 | A TMP DATA 3 | THIS IS A VERY VERY LONG STRING3 4 | THIS IS A VERY VERY LONG STRING PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD1 -5 | THIS IS A VERY VERY LONG STRING PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD1 +5 | THIS IS A VERY VERY LONG STRING PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD PAD1 PAD PAD PAD PAD ID | INFO diff --git a/test/case/result/primary-typecast.result b/test/case/result/primary-typecast.result new file mode 100644 index 00000000..36026269 --- /dev/null +++ b/test/case/result/primary-typecast.result @@ -0,0 +1,170 @@ +INITIALIZATION +CREATE TABLE TYPECAST_TABLE_1(ID INT, NAME CHAR(20), AGE FLOAT); +SUCCESS + +1. INSERT ROWS +INSERT INTO TYPECAST_TABLE_1 VALUES(666, 666, 666); +SUCCESS +INSERT INTO TYPECAST_TABLE_1 VALUES('1', '1', '1'); +SUCCESS +INSERT INTO TYPECAST_TABLE_1 VALUES('ABC', 'ABC', 'ABC'); +SUCCESS +INSERT INTO TYPECAST_TABLE_1 VALUES('0.5', '0.5', '0.5'); +SUCCESS +INSERT INTO TYPECAST_TABLE_1 VALUES('3ABC', 'ABCD', '3.14ADC'); +SUCCESS +INSERT INTO TYPECAST_TABLE_1 VALUES('4ABC', 'DCBA', 'G3.14ADC'); +SUCCESS +INSERT INTO TYPECAST_TABLE_1 VALUES(1.5, 1.5, 1.5); +SUCCESS +INSERT INTO TYPECAST_TABLE_1 VALUES(-1.5, -1.5, -1.5); +SUCCESS +SELECT * FROM TYPECAST_TABLE_1; +-2 | 1.5 | -1.5 +0 | 0.5 | 0.5 +0 | ABC | 0 +1 | 1 | 1 +2 | 1.5 | 1.5 +3 | ABCD | 3.14 +4 | DCBA | 0 +666 | 666 | 666 +ID | NAME | AGE + +2. QUERY ROWS +SELECT * FROM TYPECAST_TABLE_1 WHERE ID > 0.5; +1 | 1 | 1 +2 | 1.5 | 1.5 +3 | ABCD | 3.14 +4 | DCBA | 0 +666 | 666 | 666 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE ID > '0.1'; +1 | 1 | 1 +2 | 1.5 | 1.5 +3 | ABCD | 3.14 +4 | DCBA | 0 +666 | 666 | 666 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE ID > '1ADC'; +2 | 1.5 | 1.5 +3 | ABCD | 3.14 +4 | DCBA | 0 +666 | 666 | 666 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE ID < '1.5ADC'; +-2 | 1.5 | -1.5 +0 | 0.5 | 0.5 +0 | ABC | 0 +1 | 1 | 1 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE NAME >= 1; +-2 | 1.5 | -1.5 +1 | 1 | 1 +2 | 1.5 | 1.5 +666 | 666 | 666 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE NAME > 1.3; +-2 | 1.5 | -1.5 +2 | 1.5 | 1.5 +666 | 666 | 666 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE NAME > '1ADC'; +0 | ABC | 0 +3 | ABCD | 3.14 +4 | DCBA | 0 +666 | 666 | 666 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE AGE > 1; +2 | 1.5 | 1.5 +3 | ABCD | 3.14 +666 | 666 | 666 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE AGE > 1.2; +2 | 1.5 | 1.5 +3 | ABCD | 3.14 +666 | 666 | 666 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE AGE > 2.5; +3 | ABCD | 3.14 +666 | 666 | 666 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE AGE > '1ADC'; +2 | 1.5 | 1.5 +3 | ABCD | 3.14 +666 | 666 | 666 +ID | NAME | AGE + +3. UPDATE ROWS +UPDATE TYPECAST_TABLE_1 SET ID = 6, NAME = 6, AGE = 6 WHERE ID = 1; +SUCCESS +UPDATE TYPECAST_TABLE_1 SET ID = '66', NAME = '66', AGE = '66' WHERE ID = 2; +SUCCESS +UPDATE TYPECAST_TABLE_1 SET ID = 'JF', NAME = 'JF', AGE = 'JF' WHERE ID = -2; +SUCCESS +UPDATE TYPECAST_TABLE_1 SET ID = '0.5', NAME = '0.5', AGE = '0.5' WHERE ID = 0; +SUCCESS +UPDATE TYPECAST_TABLE_1 SET ID = '3ABC', NAME = 'ABCD', AGE = '3.14ADC' WHERE ID = 4; +SUCCESS +UPDATE TYPECAST_TABLE_1 SET ID = '4ABC', NAME = 'DCBA', AGE = 'G3.14ADC' WHERE ID = 4; +SUCCESS +UPDATE TYPECAST_TABLE_1 SET ID = 6.6, NAME = 6.6, AGE = '6.6' WHERE ID = 3; +SUCCESS +UPDATE TYPECAST_TABLE_1 SET ID = -8.8, NAME = -8.8, AGE = '-8.8' WHERE ID = 666; +SUCCESS +SELECT * FROM TYPECAST_TABLE_1; +-9 | 8.8 | -8.8 +0 | 0.5 | 0.5 +0 | 0.5 | 0.5 +0 | 0.5 | 0.5 +6 | 6 | 6 +66 | 66 | 66 +7 | 6.6 | 6.6 +7 | 6.6 | 6.6 +ID | NAME | AGE + +4. QUERY ROWS AGAIN +SELECT * FROM TYPECAST_TABLE_1 WHERE '1A' = 2; +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE '1.5A' = 2; +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE '1.5A' = 2.0; +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE '2A' = 2.0; +-9 | 8.8 | -8.8 +0 | 0.5 | 0.5 +0 | 0.5 | 0.5 +0 | 0.5 | 0.5 +6 | 6 | 6 +66 | 66 | 66 +7 | 6.6 | 6.6 +7 | 6.6 | 6.6 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE '2.0-A' = 2.0; +-9 | 8.8 | -8.8 +0 | 0.5 | 0.5 +0 | 0.5 | 0.5 +0 | 0.5 | 0.5 +6 | 6 | 6 +66 | 66 | 66 +7 | 6.6 | 6.6 +7 | 6.6 | 6.6 +ID | NAME | AGE + +5. TYPE PROMOTION +SELECT * FROM TYPECAST_TABLE_1 WHERE AGE >= 7.0; +66 | 66 | 66 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE AGE >= 7; +66 | 66 | 66 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE ID >= 6; +6 | 6 | 6 +66 | 66 | 66 +7 | 6.6 | 6.6 +7 | 6.6 | 6.6 +ID | NAME | AGE +SELECT * FROM TYPECAST_TABLE_1 WHERE ID >= 6.25; +66 | 66 | 66 +7 | 6.6 | 6.6 +7 | 6.6 | 6.6 +ID | NAME | AGE diff --git a/test/case/result/primary-unique.result b/test/case/result/primary-unique.result index 54a36709..e8810115 100644 --- a/test/case/result/primary-unique.result +++ b/test/case/result/primary-unique.result @@ -1,24 +1,47 @@ INITIALIZATION -CREATE TABLE unique_table(id int, col1 int, col2 int); +CREATE TABLE UNIQUE_TABLE(ID INT, COL1 INT, COL2 INT); SUCCESS -INSERT INTO unique_table VALUES (1,1,1); +CREATE TABLE UNIQUE_TABLE2(ID1 INT, ID2 INT, COL1 INT, COL2 INT); +SUCCESS +INSERT INTO UNIQUE_TABLE VALUES (1,1,1); +SUCCESS +INSERT INTO UNIQUE_TABLE2 VALUES (1,1,1,1); SUCCESS 1. UNIQUE TEST -CREATE UNIQUE INDEX index_id on unique_table(id); +CREATE UNIQUE INDEX INDEX_ID ON UNIQUE_TABLE(ID); SUCCESS -INSERT INTO unique_table VALUES (2,1,1); +INSERT INTO UNIQUE_TABLE VALUES (2,1,1); SUCCESS -CREATE UNIQUE INDEX index_id on unique_table(id); +CREATE UNIQUE INDEX INDEX_ID ON UNIQUE_TABLE(ID); FAILURE -INSERT INTO unique_table VALUES (3,2,1); +INSERT INTO UNIQUE_TABLE VALUES (3,2,1); +SUCCESS +INSERT INTO UNIQUE_TABLE VALUES (1,2,1); +FAILURE + +CREATE UNIQUE INDEX INDEX_ID ON UNIQUE_TABLE2(ID1, ID2); SUCCESS -INSERT INTO unique_table VALUES (1,2,1); +INSERT INTO UNIQUE_TABLE2 VALUES (1,1,2,1); FAILURE +INSERT INTO UNIQUE_TABLE2 VALUES (1,2,1,1); +SUCCESS +INSERT INTO UNIQUE_TABLE2 VALUES (1,2,2,1); +FAILURE +INSERT INTO UNIQUE_TABLE2 VALUES (1,3,4,2); +SUCCESS +INSERT INTO UNIQUE_TABLE2 VALUES (2,1,1,1); +SUCCESS 2. SELECT -SELECT * FROM unique_table; +SELECT * FROM UNIQUE_TABLE; 1 | 1 | 1 2 | 1 | 1 3 | 2 | 1 ID | COL1 | COL2 +SELECT * FROM UNIQUE_TABLE2; +1 | 1 | 1 | 1 +1 | 2 | 1 | 1 +1 | 3 | 4 | 2 +2 | 1 | 1 | 1 +ID1 | ID2 | COL1 | COL2 diff --git a/test/case/result/primary-update-plus.result b/test/case/result/primary-update-plus.result new file mode 100644 index 00000000..9c38bf76 --- /dev/null +++ b/test/case/result/primary-update-plus.result @@ -0,0 +1,71 @@ +INITIALIZATION +CREATE TABLE STUDENT (ID INT, NAME CHAR(9), MAJOR CHAR(32)); +SUCCESS +INSERT INTO STUDENT VALUES (0, 'KANGKANGA', 'AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA'); +SUCCESS +INSERT INTO STUDENT VALUES (1, 'TOMTOMTOM', 'COMPUTER SCIENCECOMPUTER SCIENCE'); +SUCCESS +INSERT INTO STUDENT VALUES (2, 'JERRYJERR', 'COMPUTER SCIENCECOMPUTER SCIENCE'); +SUCCESS +INSERT INTO STUDENT VALUES (3, 'JACKJACKJ', 'ELECTRICAL ENGINEERINGER SCIENCE'); +SUCCESS +INSERT INTO STUDENT VALUES (3, 'JERRYJERR', 'AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA'); +SUCCESS + +1. UPDATE A ROW +UPDATE STUDENT SET MAJOR = 'ELECTRICAL ENGINEERING' WHERE ID = 2; +SUCCESS +SELECT * FROM STUDENT WHERE ID >= 1; +1 | TOMTOMTOM | COMPUTER SCIENCECOMPUTER SCIENCE +2 | JERRYJERR | ELECTRICAL ENGINEERING +3 | JACKJACKJ | ELECTRICAL ENGINEERINGER SCIENCE +3 | JERRYJERR | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA +ID | NAME | MAJOR +SELECT * FROM STUDENT WHERE ID = 0; +0 | KANGKANGA | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA +ID | NAME | MAJOR + +2. UPDATE NON-EXISTENT ROW +UPDATE STUDENT SET ID = 100 WHERE NAME = 'JERRYJERRY'; +SUCCESS +SELECT * FROM STUDENT WHERE ID > 2; +3 | JACKJACKJ | ELECTRICAL ENGINEERINGER SCIENCE +3 | JERRYJERR | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA +ID | NAME | MAJOR +SELECT * FROM STUDENT WHERE ID < 101; +0 | KANGKANGA | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA +1 | TOMTOMTOM | COMPUTER SCIENCECOMPUTER SCIENCE +2 | JERRYJERR | ELECTRICAL ENGINEERING +3 | JACKJACKJ | ELECTRICAL ENGINEERINGER SCIENCE +3 | JERRYJERR | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA +ID | NAME | MAJOR + +3. UPDATE ROWS +UPDATE STUDENT SET NAME = '12345678' WHERE ID < 3; +SUCCESS +SELECT * FROM STUDENT WHERE ID = 2; +2 | 12345678 | ELECTRICAL ENGINEERING +ID | NAME | MAJOR +SELECT * FROM STUDENT WHERE NAME = '12345678'; +0 | 12345678 | AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA +1 | 12345678 | COMPUTER SCIENCECOMPUTER SCIENCE +2 | 12345678 | ELECTRICAL ENGINEERING +ID | NAME | MAJOR + +4. UPDATE ROWS AGAIN +UPDATE STUDENT SET NAME = '9876543210' WHERE ID < 3; +FAILURE +SELECT * FROM STUDENT WHERE ID = 2; +2 | 12345678 | ELECTRICAL ENGINEERING +ID | NAME | MAJOR +SELECT * FROM STUDENT WHERE NAME = '9876543210'; +ID | NAME | MAJOR + +5. UPDATE MULTIPLE COLUMNS +UPDATE STUDENT SET NAME = 'MULNAME', ID = 666, MAJOR = 'MULMAJOR' WHERE ID = 1; +SUCCESS +SELECT * FROM STUDENT WHERE ID = 1; +ID | NAME | MAJOR +SELECT * FROM STUDENT WHERE NAME = 'MULNAME' AND MAJOR = 'MULMAJOR'; +666 | MULNAME | MULMAJOR +ID | NAME | MAJOR diff --git a/test/case/result/primary-update-select.result b/test/case/result/primary-update-select.result new file mode 100644 index 00000000..1bdc4267 --- /dev/null +++ b/test/case/result/primary-update-select.result @@ -0,0 +1,126 @@ +INIT DATA +CREATE TABLE UPDATE_TABLE_2(ID INT, T_NAME CHAR(4), COL1 INT, COL2 INT); +SUCCESS +CREATE TABLE UPDATE_TABLE_3(ID INT, T_NAME CHAR(4), COL1 INT, COL2 INT); +SUCCESS +INSERT INTO UPDATE_TABLE_2 VALUES (1,'N1',1,1); +SUCCESS +INSERT INTO UPDATE_TABLE_2 VALUES (2,'N2',1,1); +SUCCESS +INSERT INTO UPDATE_TABLE_2 VALUES (3,'N3',3,1); +SUCCESS +INSERT INTO UPDATE_TABLE_2 VALUES (4,'N3',4,2); +SUCCESS +INSERT INTO UPDATE_TABLE_2 VALUES (5,'N3',5,3); +SUCCESS +INSERT INTO UPDATE_TABLE_3 VALUES (1,'N1',1,1); +SUCCESS +INSERT INTO UPDATE_TABLE_3 VALUES (2,'N2',2,2); +SUCCESS +INSERT INTO UPDATE_TABLE_3 VALUES (3,'N3',3,3); +SUCCESS +INSERT INTO UPDATE_TABLE_3 VALUES (4,'N4',4,4); +SUCCESS +INSERT INTO UPDATE_TABLE_3 VALUES (5,'N5',5,5); +SUCCESS +INSERT INTO UPDATE_TABLE_3 VALUES (6,'N6',6,6); +SUCCESS +INSERT INTO UPDATE_TABLE_3 VALUES (6,'N6',6,6); +SUCCESS + +1. UPDATE-SELECT +UPDATE UPDATE_TABLE_2 SET T_NAME='N01', COL2=2 WHERE ID=1; +SUCCESS +UPDATE UPDATE_TABLE_2 SET COL1=2, COL2=(SELECT UPDATE_TABLE_3.COL2 FROM UPDATE_TABLE_3 WHERE UPDATE_TABLE_3.ID=1) WHERE ID=1; +SUCCESS +SELECT * FROM UPDATE_TABLE_2; +1 | N01 | 2 | 1 +2 | N2 | 1 | 1 +3 | N3 | 3 | 1 +4 | N3 | 4 | 2 +5 | N3 | 5 | 3 +ID | T_NAME | COL1 | COL2 + +2. UPDATE-SELECT ROWS +UPDATE UPDATE_TABLE_2 SET COL1=(SELECT UPDATE_TABLE_3.COL2 FROM UPDATE_TABLE_3) WHERE ID=1; +FAILURE +UPDATE UPDATE_TABLE_2 SET COL1=(SELECT UPDATE_TABLE_3.COL2 FROM UPDATE_TABLE_3 WHERE UPDATE_TABLE_3.ID>1) WHERE ID=1; +FAILURE +UPDATE UPDATE_TABLE_2 SET T_NAME=(SELECT UPDATE_TABLE_3.COL2 FROM UPDATE_TABLE_3 WHERE UPDATE_TABLE_3.ID<1) WHERE ID=1; +SUCCESS +SELECT * FROM UPDATE_TABLE_2; +1 | NULL | 2 | 1 +2 | N2 | 1 | 1 +3 | N3 | 3 | 1 +4 | N3 | 4 | 2 +5 | N3 | 5 | 3 +ID | T_NAME | COL1 | COL2 + +3. UPDATE ROWS-SELECT ROWS +UPDATE UPDATE_TABLE_3 SET COL1=(SELECT COL2 FROM UPDATE_TABLE_2); +FAILURE +SELECT * FROM UPDATE_TABLE_3; +1 | N1 | 1 | 1 +2 | N2 | 2 | 2 +3 | N3 | 3 | 3 +4 | N4 | 4 | 4 +5 | N5 | 5 | 5 +6 | N6 | 6 | 6 +6 | N6 | 6 | 6 +ID | T_NAME | COL1 | COL2 + +3. UPDATE ROWS-SELECT COLS AND ROWS +UPDATE UPDATE_TABLE_3 SET COL1=(SELECT COL1, COL2 FROM UPDATE_TABLE_2); +FAILURE +SELECT * FROM UPDATE_TABLE_3; +1 | N1 | 1 | 1 +2 | N2 | 2 | 2 +3 | N3 | 3 | 3 +4 | N4 | 4 | 4 +5 | N5 | 5 | 5 +6 | N6 | 6 | 6 +6 | N6 | 6 | 6 +ID | T_NAME | COL1 | COL2 + +4. CHECK TYPECAST +UPDATE UPDATE_TABLE_3 SET T_NAME=813,COL1=547.82 WHERE ID=2; +SUCCESS +UPDATE UPDATE_TABLE_3 SET T_NAME=(SELECT UPDATE_TABLE_2.COL1 FROM UPDATE_TABLE_2 WHERE UPDATE_TABLE_2.ID=3), COL2=(SELECT MIN(UPDATE_TABLE_3.COL2) FROM UPDATE_TABLE_3 WHERE UPDATE_TABLE_3.ID=1) WHERE ID=5; +SUCCESS +SELECT * FROM UPDATE_TABLE_3; +1 | N1 | 1 | 1 +2 | 813 | 548 | 2 +3 | N3 | 3 | 3 +4 | N4 | 4 | 4 +5 | 3 | 5 | 1 +6 | N6 | 6 | 6 +6 | N6 | 6 | 6 +ID | T_NAME | COL1 | COL2 + +5. SUBQUERY RETURN NULL +UPDATE UPDATE_TABLE_3 SET COL1=2, COL2=(SELECT UPDATE_TABLE_2.COL2 FROM UPDATE_TABLE_2 WHERE UPDATE_TABLE_2.ID=999) WHERE ID=4; +SUCCESS +UPDATE UPDATE_TABLE_3 SET COL1=2, COL2=(SELECT UPDATE_TABLE_2.COL2 FROM UPDATE_TABLE_2 WHERE UPDATE_TABLE_2.ID=999) WHERE ID=666; +SUCCESS +SELECT * FROM UPDATE_TABLE_3; +1 | N1 | 1 | 1 +2 | 813 | 548 | 2 +3 | N3 | 3 | 3 +4 | N4 | 2 | NULL +5 | 3 | 5 | 1 +6 | N6 | 6 | 6 +6 | N6 | 6 | 6 +ID | T_NAME | COL1 | COL2 + +6. SUBQUERY RETURN MULTIPLE ROWS +UPDATE UPDATE_TABLE_2 SET T_NAME=(SELECT UPDATE_TABLE_3.T_NAME FROM UPDATE_TABLE_3 WHERE UPDATE_TABLE_3.ID=6) WHERE COL1=5 AND COL2=3; +FAILURE +UPDATE UPDATE_TABLE_2 SET T_NAME=(SELECT UPDATE_TABLE_3.T_NAME FROM UPDATE_TABLE_3 WHERE UPDATE_TABLE_3.ID=6) WHERE COL1=5 AND COL2=6; +SUCCESS +SELECT * FROM UPDATE_TABLE_2; +1 | NULL | 2 | 1 +2 | N2 | 1 | 1 +3 | N3 | 3 | 1 +4 | N3 | 4 | 2 +5 | N3 | 5 | 3 +ID | T_NAME | COL1 | COL2 diff --git a/test/case/result/primary-update.result b/test/case/result/primary-update.result index 86b2069c..6c376314 100644 --- a/test/case/result/primary-update.result +++ b/test/case/result/primary-update.result @@ -1,5 +1,5 @@ INITIALIZATION -CREATE TABLE Update_table_1(id int, t_name char, col1 int, col2 int); +CREATE TABLE Update_table_1(id int, t_name char(4), col1 int, col2 int); SUCCESS CREATE INDEX index_id on Update_table_1(id); SUCCESS @@ -78,4 +78,4 @@ ID | T_NAME | COL1 | COL2 10. UPDATE WITH INVALID VALUE UPDATE Update_table_1 SET col1='N01' WHERE id=1; -FAILURE +SUCCESS diff --git a/test/case/test/basic.test b/test/case/test/basic.test index 5d7ddd02..dab4ccd2 100644 --- a/test/case/test/basic.test +++ b/test/case/test/basic.test @@ -48,4 +48,4 @@ calc (1+2) * (2 * (20+-(5*1))); calc 20/0; calc 13.2/0.0; calc "123" + 4; -calc "abc" + 4; \ No newline at end of file +calc "abc" + 4; diff --git a/test/case/test/primary-aggregation-func.test b/test/case/test/primary-aggregation-func.test index 478c3136..58dfe82c 100644 --- a/test/case/test/primary-aggregation-func.test +++ b/test/case/test/primary-aggregation-func.test @@ -1,5 +1,5 @@ -- echo initialization -CREATE TABLE aggregation_func(id int, num int, price float, addr char, birthday date); +CREATE TABLE aggregation_func(id int, num int, price float, addr char(4), birthday date); INSERT INTO aggregation_func VALUES (1, 18, 10.0, 'abc', '2020-01-01'); INSERT INTO aggregation_func VALUES (2, 15, 20.0, 'abc', '2010-01-11'); diff --git a/test/case/test/primary-complex-sub-query-plus.test b/test/case/test/primary-complex-sub-query-plus.test new file mode 100644 index 00000000..12dad880 --- /dev/null +++ b/test/case/test/primary-complex-sub-query-plus.test @@ -0,0 +1,31 @@ +-- echo initialization +CREATE TABLE csq_1(id int, col1 int, feat1 float); +CREATE TABLE csq_2(id int, col2 int, feat2 float); +CREATE TABLE csq_3(id int, col3 int, feat3 float); + +INSERT INTO csq_1 VALUES (1, 4, 11.2); +INSERT INTO csq_1 VALUES (2, 2, 12.0); +INSERT INTO csq_1 VALUES (3, 3, 13.5); +INSERT INTO csq_1 VALUES (4, 1, 10.8); +INSERT INTO csq_1 VALUES (5, 5, 15.3); +INSERT INTO csq_1 VALUES (6, 7, 12.4); +INSERT INTO csq_1 VALUES (7, 6, 11.9); + +INSERT INTO csq_2 VALUES (1, 2, 13.0); +INSERT INTO csq_2 VALUES (2, 7, 10.5); +INSERT INTO csq_2 VALUES (3, 9, 13.8); +INSERT INTO csq_2 VALUES (4, 5, 14.1); +INSERT INTO csq_2 VALUES (5, 3, 12.6); +INSERT INTO csq_2 VALUES (6, 1, 10.2); +INSERT INTO csq_2 VALUES (7, 4, 12.3); + +INSERT INTO csq_3 VALUES (1, 2, 11.0); +INSERT INTO csq_3 VALUES (2, 3, 13.4); +INSERT INTO csq_3 VALUES (3, 6, 16.5); +INSERT INTO csq_3 VALUES (4, 4, 12.9); +INSERT INTO csq_3 VALUES (5, 5, 14.6); +INSERT INTO csq_3 VALUES (6, 7, 11.7); +INSERT INTO csq_3 VALUES (7, 8, 15.0); + +-- echo 1. select +-- sort select * from csq_1 where feat1 <> (select min(csq_2.feat2) from csq_2 where csq_2.feat2 > csq_1.feat1); diff --git a/test/case/test/primary-complex-sub-query.test b/test/case/test/primary-complex-sub-query.test index e5ef032b..ded38b1a 100644 --- a/test/case/test/primary-complex-sub-query.test +++ b/test/case/test/primary-complex-sub-query.test @@ -37,6 +37,8 @@ INSERT INTO csq_3 VALUES (5, 5, 14.6); -- sort select * from csq_1 where feat1 <> (select avg(csq_2.feat2) from csq_2 where csq_2.feat2 > csq_1.feat1); +-- sort select * from csq_1 where feat1 > (select min(csq_2.feat2) from csq_2 where csq_2.feat2 > csq_1.feat1); + -- sort select * from csq_1 where col1 not in (select csq_2.col2 from csq_2 where csq_2.id in (select csq_3.id from csq_3 where csq_1.id = csq_3.id)); -- echo 2. Select with empty table diff --git a/test/case/test/primary-create-view.test b/test/case/test/primary-create-view.test new file mode 100644 index 00000000..e69de29b diff --git a/test/case/test/primary-drop-table.test b/test/case/test/primary-drop-table.test index 8d15a47a..5fbd448d 100644 --- a/test/case/test/primary-drop-table.test +++ b/test/case/test/primary-drop-table.test @@ -1,37 +1,37 @@ -- echo 1. Drop empty table -CREATE TABLE Drop_table_1(id int, t_name char); +CREATE TABLE Drop_table_1(id int, t_name char(4)); DROP TABLE Drop_table_1; -- echo 2. Drop non-empty table -CREATE TABLE Drop_table_2(id int, t_name char); +CREATE TABLE Drop_table_2(id int, t_name char(4)); INSERT INTO Drop_table_2 VALUES (1,'OB'); DROP TABLE Drop_table_2; -- echo 3. Check the accuracy of dropping table -CREATE TABLE Drop_table_3(id int, t_name char); +CREATE TABLE Drop_table_3(id int, t_name char(4)); INSERT INTO Drop_table_3 VALUES (1,'OB'); -- sort SELECT * FROM Drop_table_3; DROP TABLE Drop_table_3; INSERT INTO Drop_table_3 VALUES (1,'OB'); SELECT * FROM Drop_table_3; DELETE FROM Drop_table_3 WHERE id = 3; -CREATE TABLE Drop_table_3(id int, t_name char); +CREATE TABLE Drop_table_3(id int, t_name char(4)); -- sort SELECT * FROM Drop_table_3; -- echo 4. Drop non-existent table -CREATE TABLE Drop_table_4(id int, t_name char); +CREATE TABLE Drop_table_4(id int, t_name char(4)); DROP TABLE Drop_table_4; DROP TABLE Drop_table_4; DROP TABLE Drop_table_4_1; -- echo 5. Create a table which has dropped -CREATE TABLE Drop_table_5(id int, t_name char); +CREATE TABLE Drop_table_5(id int, t_name char(4)); DROP TABLE Drop_table_5; -CREATE TABLE Drop_table_5(id int, t_name char); +CREATE TABLE Drop_table_5(id int, t_name char(4)); SELECT * FROM Drop_table_5; -- echo 6. Drop a table with index -CREATE TABLE Drop_table_6(id int, t_name char); +CREATE TABLE Drop_table_6(id int, t_name char(4)); CREATE INDEX index_id on Drop_table_6(id); INSERT INTO Drop_table_6 VALUES (1,'OB'); -- sort SELECT * FROM Drop_table_6; diff --git a/test/case/test/primary-function.test b/test/case/test/primary-function.test new file mode 100644 index 00000000..4df29df0 --- /dev/null +++ b/test/case/test/primary-function.test @@ -0,0 +1,16 @@ +-- echo 1. init data +CREATE TABLE function_table(id int, name char(20), score float, u_date date); +CREATE TABLE function_table_2(id int, name char(30), score float, u_date date); +insert into function_table VALUES (4, 'KI83C4RH0AGX9', 7.61, '2017-02-06'); +insert into function_table VALUES (4, 'FXMB2DOJ4', 8.35, '2021-10-07'); +insert into function_table VALUES (8, 'Q3NS46CJ', 8.66, '2023-06-19'); +insert into function_table VALUES (1, 'Q3NS4', 4.90, '2023-06-19'); +insert into function_table VALUES (3, 'Q3NS46CJ012311234561', 6.50, '2023-06-19'); +insert into function_table VALUES (2, 'Q3CJ', 9.00, '2023-06-19'); + +-- echo 2. invalid date +-- sort select id, date_format(name, '%y-%m-%d') from function_table; + +-- echo 3. round test +-- sort select id, length(name), round(score) from function_table where id<8; +-- sort select * from function_table; diff --git a/test/case/test/primary-group-by.test b/test/case/test/primary-group-by.test index 01836179..23e82e87 100644 --- a/test/case/test/primary-group-by.test +++ b/test/case/test/primary-group-by.test @@ -1,6 +1,6 @@ -- echo 1. create table -create table t_group_by (id int, score float, name char); -create table t_group_by_2 (id int, age int); +create table t_group_by (id int not null, score float not null, name char(1) null); +create table t_group_by_2 (id int not null, age int not null); -- echo 2. insert records insert into t_group_by values(3, 1.0, 'a'); @@ -10,6 +10,7 @@ insert into t_group_by values(3, 2.0, 'c'); insert into t_group_by values(3, 4.0, 'c'); insert into t_group_by values(3, 3.0, 'd'); insert into t_group_by values(3, 2.0, 'f'); +insert into t_group_by values(4, 3.0, null); insert into t_group_by_2 values(1, 10); insert into t_group_by_2 values(2, 20); @@ -21,6 +22,8 @@ insert into t_group_by_2 values(4, 20); -- echo 3. primary group by -- sort select id, avg(score) from t_group_by group by id; +-- sort select id, sum(score) from t_group_by group by id; + -- sort select name, min(id), max(score) from t_group_by group by name; -- sort select id, name, avg(score) from t_group_by group by id, name; @@ -31,4 +34,7 @@ insert into t_group_by_2 values(4, 20); -- sort select name, count(id), max(score) from t_group_by where name > 'a' and id>=0 group by name; -- echo 5. multi table --- sort select t_group_by.id, t_group_by.name, avg(t_group_by.score), avg(t_group_by_2.age) from t_group_by, t_group_by_2 where t_group_by.id=t_group_by_2.id group by t_group_by.id, t_group_by.name; \ No newline at end of file +-- sort select t_group_by.id, t_group_by.name, avg(t_group_by.score), avg(t_group_by_2.age) from t_group_by, t_group_by_2 where t_group_by.id=t_group_by_2.id group by t_group_by.id, t_group_by.name; + +-- echo 6. having typecast +-- sort select id, avg(score) from t_group_by group by id having avg(score)>2; diff --git a/test/case/test/primary-insert.test b/test/case/test/primary-insert.test index 429afc87..96c831f2 100644 --- a/test/case/test/primary-insert.test +++ b/test/case/test/primary-insert.test @@ -1,13 +1,27 @@ -- echo initialization -CREATE TABLE insert_table(id int, t_name char, col1 int, col2 int); +CREATE TABLE insert_table(id int, t_name char(4), col1 int, col2 int); +CREATE TABLE insert_table2(id int, t_name char(6), col1 int, col2 int); -- echo 1. insert INSERT INTO insert_table VALUES (1,'N1',1,1); -INSERT INTO insert_table VALUES (2,'N2',1,1),(3,'N3',2,1); +INSERT INTO insert_table VALUES (1,'N123',1,1); +INSERT INTO insert_table VALUES (2,'N123',1,1),(3,'N32',2,1); + +INSERT INTO insert_table2 VALUES (1,'N1',1,1); +INSERT INTO insert_table2 VALUES (1,'N123',1,1); +INSERT INTO insert_table2 VALUES (2,'N1233',1,1),(3,'N3211',2,1); +INSERT INTO insert_table2 VALUES (4,'','',''); -- echo 2. error -INSERT INTO insert_table VALUES (4,'N4',1,1),(1,1,1); -INSERT INTO insert_table VALUES (4,'N4',1,1),(1,1,1,1); +INSERT INTO insert_table VALUES (5,'N4',1,1),(1,1,1); +INSERT INTO insert_table VALUES (5,'N4',1,1),(1,1,1,1,1); +INSERT INTO insert_table VALUES (5,'12345',1,1); + +INSERT INTO insert_table2 VALUES (5,'N412',1,1),(1,1,1); +INSERT INTO insert_table2 VALUES (5,'N41',1,1),(1,1,1,1,1); +INSERT INTO insert_table2 VALUES (5,'1234567',1,1); +INSERT INTO insert_table2 VALUES (5,'12345678910',1,1); -- echo 3. select --- sort SELECT * FROM insert_table; \ No newline at end of file +-- sort SELECT * FROM insert_table; +-- sort SELECT * FROM insert_table2; diff --git a/test/case/test/primary-null.test b/test/case/test/primary-null.test index e86fead7..61fb6a10 100644 --- a/test/case/test/primary-null.test +++ b/test/case/test/primary-null.test @@ -1,6 +1,6 @@ -- echo initialization -CREATE TABLE null_table(id int, num int nullable, price float not null, birthday date nullable); -CREATE TABLE null_table2(id int, num int nullable, price float not null, birthday date nullable); +CREATE TABLE null_table(id int not null, num int, price float not null, birthday date null); +CREATE TABLE null_table2(id int not null, num int, price float not null, birthday date null); CREATE INDEX index_num on null_table(num); -- echo 1. insert @@ -74,7 +74,7 @@ SELECT count(birthday) FROM null_table; SELECT avg(num) FROM null_table; -- echo 6. aggregation with null columns -CREATE TABLE null_table3(id int, num int nullable); +CREATE TABLE null_table3(id int not null, num int null); INSERT INTO null_table3 VALUES (1, null); INSERT INTO null_table3 VALUES (2, null); SELECT count(num) FROM null_table3; diff --git a/test/case/test/primary-show-index.test b/test/case/test/primary-show-index.test new file mode 100644 index 00000000..46ca6220 --- /dev/null +++ b/test/case/test/primary-show-index.test @@ -0,0 +1,19 @@ +-- echo INITIALIZATION +CREATE TABLE Index_table_1(id int, age int); +insert into Index_table_1 values (1,1); +insert into Index_table_1 values (2,2); +insert into Index_table_1 values (3,3); +insert into Index_table_1 values (1,2); +insert into Index_table_1 values (1,3); + +-- echo 1. show empty index +SHOW INDEX FROM Index_table_1; + +-- echo 2. show an index +CREATE INDEX index_id_1 on Index_table_1(id); +SHOW INDEX FROM Index_table_1; + +-- echo 3. show indexes +CREATE INDEX index_id_2 on Index_table_1(id); +CREATE INDEX index_id_3 on Index_table_1(age); +SHOW INDEX FROM Index_table_1; diff --git a/test/case/test/primary-simple-sub-query.test b/test/case/test/primary-simple-sub-query.test index 7ae8478b..1145c462 100644 --- a/test/case/test/primary-simple-sub-query.test +++ b/test/case/test/primary-simple-sub-query.test @@ -32,15 +32,20 @@ INSERT INTO ssq_2 VALUES (5, 3, 12.6); -- sort select * from ssq_1 where feat1 <> (select avg(ssq_2.feat2) from ssq_2); -- echo 2. Select with empty table +-- sort select * from ssq_1 where col1 not in (2,null); -- sort select * from ssq_1 where feat1 < (select max(ssq_2.feat2) from ssq_2 where 1=0); -- sort select * from ssq_1 where id in (select ssq_2.id from ssq_2 where 1=0); -- sort select * from ssq_1 where id not in (select ssq_2.id from ssq_2 where 1=0); -- sort select * from ssq_3 where feat3 < (select max(ssq_2.feat2) from ssq_2); -- sort select * from ssq_3 where id in (select ssq_2.id from ssq_2); -- sort select * from ssq_3 where id not in (select ssq_2.id from ssq_2); +-- sort select * from ssq_1 where (select ssq_2.id from ssq_2 where col2 = 52) = id; ---echo 3. error +-- echo 3. error select * from ssq_1 where col1 = (select ssq_2.col2 from ssq_2); select * from ssq_1 where col1 = (select * from ssq_2); select * from ssq_1 where col1 in (select * from ssq_2); -select * from ssq_1 where col1 not in (select * from ssq_2); \ No newline at end of file +select * from ssq_1 where col1 not in (select * from ssq_2); + +-- echo 4. stream clear after error +-- sort select * from ssq_1 where (select ssq_2.id from ssq_2 where col2 = 52) = id; diff --git a/test/case/test/primary-typecast.test b/test/case/test/primary-typecast.test new file mode 100644 index 00000000..b4dd2ed9 --- /dev/null +++ b/test/case/test/primary-typecast.test @@ -0,0 +1,50 @@ +-- echo initialization +CREATE TABLE Typecast_table_1(id int, name char(20), age float); + +-- echo 1. insert rows +INSERT INTO Typecast_table_1 VALUES(666, 666, 666); +INSERT INTO Typecast_table_1 VALUES('1', '1', '1'); +INSERT INTO Typecast_table_1 VALUES('abc', 'abc', 'abc'); +INSERT INTO Typecast_table_1 VALUES('0.5', '0.5', '0.5'); +INSERT INTO Typecast_table_1 VALUES('3abc', 'abcd', '3.14adc'); +INSERT INTO Typecast_table_1 VALUES('4abc', 'dcba', 'g3.14adc'); +INSERT INTO Typecast_table_1 VALUES(1.5, 1.5, 1.5); +INSERT INTO Typecast_table_1 VALUES(-1.5, -1.5, -1.5); +-- sort select * from Typecast_table_1; + +-- echo 2. query rows +-- sort SELECT * from Typecast_table_1 WHERE id > 0.5; +-- sort SELECT * from Typecast_table_1 WHERE id > '0.1'; +-- sort SELECT * from Typecast_table_1 WHERE id > '1adc'; +-- sort SELECT * from Typecast_table_1 WHERE id < '1.5adc'; +-- sort SELECT * from Typecast_table_1 WHERE name >= 1; +-- sort SELECT * from Typecast_table_1 WHERE name > 1.3; +-- sort SELECT * from Typecast_table_1 WHERE name > '1adc'; +-- sort SELECT * from Typecast_table_1 WHERE age > 1; +-- sort SELECT * from Typecast_table_1 WHERE age > 1.2; +-- sort SELECT * from Typecast_table_1 WHERE age > 2.5; +-- sort SELECT * from Typecast_table_1 WHERE age > '1adc'; + +-- echo 3. update rows +UPDATE Typecast_table_1 SET id = 6, name = 6, age = 6 where id = 1; +UPDATE Typecast_table_1 SET id = '66', name = '66', age = '66' where id = 2; +UPDATE Typecast_table_1 SET id = 'jf', name = 'jf', age = 'jf' where id = -2; +UPDATE Typecast_table_1 SET id = '0.5', name = '0.5', age = '0.5' where id = 0; +UPDATE Typecast_table_1 SET id = '3abc', name = 'abcd', age = '3.14adc' where id = 4; +UPDATE Typecast_table_1 SET id = '4abc', name = 'dcba', age = 'g3.14adc' where id = 4; +UPDATE Typecast_table_1 SET id = 6.6, name = 6.6, age = '6.6' where id = 3; +UPDATE Typecast_table_1 SET id = -8.8, name = -8.8, age = '-8.8' where id = 666; +-- sort select * from Typecast_table_1; + +-- echo 4. query rows again +-- sort SELECT * FROM Typecast_table_1 WHERE '1a' = 2; +-- sort SELECT * FROM Typecast_table_1 WHERE '1.5a' = 2; +-- sort SELECT * FROM Typecast_table_1 WHERE '1.5a' = 2.0; +-- sort SELECT * FROM Typecast_table_1 WHERE '2a' = 2.0; +-- sort SELECT * FROM Typecast_table_1 WHERE '2.0-a' = 2.0; + +-- echo 5. type promotion +-- sort SELECT * FROM Typecast_table_1 WHERE age >= 7.0; +-- sort SELECT * FROM Typecast_table_1 WHERE age >= 7; +-- sort SELECT * FROM Typecast_table_1 WHERE id >= 6; +-- sort SELECT * FROM Typecast_table_1 WHERE id >= 6.25; diff --git a/test/case/test/primary-unique.test b/test/case/test/primary-unique.test index 7ced3ef4..0c3b18b1 100644 --- a/test/case/test/primary-unique.test +++ b/test/case/test/primary-unique.test @@ -1,6 +1,8 @@ -- echo initialization CREATE TABLE unique_table(id int, col1 int, col2 int); +CREATE TABLE unique_table2(id1 int, id2 int, col1 int, col2 int); INSERT INTO unique_table VALUES (1,1,1); +INSERT INTO unique_table2 VALUES (1,1,1,1); -- echo 1. unique test CREATE UNIQUE INDEX index_id on unique_table(id); @@ -9,5 +11,13 @@ CREATE UNIQUE INDEX index_id on unique_table(id); INSERT INTO unique_table VALUES (3,2,1); INSERT INTO unique_table VALUES (1,2,1); +CREATE UNIQUE INDEX index_id on unique_table2(id1, id2); +INSERT INTO unique_table2 VALUES (1,1,2,1); +INSERT INTO unique_table2 VALUES (1,2,1,1); +INSERT INTO unique_table2 VALUES (1,2,2,1); +INSERT INTO unique_table2 VALUES (1,3,4,2); +INSERT INTO unique_table2 VALUES (2,1,1,1); + -- echo 2. select --- sort SELECT * FROM unique_table; \ No newline at end of file +-- sort SELECT * FROM unique_table; +-- sort SELECT * FROM unique_table2; diff --git a/test/case/test/primary-update-plus.test b/test/case/test/primary-update-plus.test new file mode 100644 index 00000000..bc4e49a1 --- /dev/null +++ b/test/case/test/primary-update-plus.test @@ -0,0 +1,32 @@ +-- echo initialization +create table student (id int, name char(9), major char(32)); +insert into student values (0, 'KangKanga', 'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa'); +insert into student values (1, 'TomTomTom', 'Computer ScienceComputer Science'); +insert into student values (2, 'JerryJerr', 'Computer ScienceComputer Science'); +insert into student values (3, 'JackJackJ', 'Electrical Engineeringer Science'); +insert into student values (3, 'JerryJerr', 'aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa'); + +-- echo 1. update a row +update student set major = 'Electrical Engineering' where id = 2; +-- sort select * from student where id >= 1; +-- sort select * from student where id = 0; + +-- echo 2. update non-existent row +update student set id = 100 where name = 'JerryJerry'; +-- sort select * from student where id > 2; +-- sort select * from student where id < 101; + +-- echo 3. update rows +update student set name = '12345678' where id < 3; +-- sort select * from student where id = 2; +-- sort select * from student where name = '12345678'; + +-- echo 4. update rows again +update student set name = '9876543210' where id < 3; +-- sort select * from student where id = 2; +-- sort select * from student where name = '9876543210'; + +-- echo 5. update multiple columns +update student set name = 'MulName', id = 666, major = 'MulMajor' where id = 1; +-- sort select * from student where id = 1; +-- sort select * from student where name = 'MulName' and major = 'MulMajor'; diff --git a/test/case/test/primary-update-select.test b/test/case/test/primary-update-select.test new file mode 100644 index 00000000..6733cc03 --- /dev/null +++ b/test/case/test/primary-update-select.test @@ -0,0 +1,49 @@ +-- echo init data +CREATE TABLE Update_table_2(id int, t_name char(4), col1 int, col2 int); +CREATE TABLE Update_table_3(id int, t_name char(4), col1 int, col2 int); +INSERT INTO Update_table_2 VALUES (1,'N1',1,1); +INSERT INTO Update_table_2 VALUES (2,'N2',1,1); +INSERT INTO Update_table_2 VALUES (3,'N3',3,1); +INSERT INTO Update_table_2 VALUES (4,'N3',4,2); +INSERT INTO Update_table_2 VALUES (5,'N3',5,3); +INSERT INTO Update_table_3 VALUES (1,'N1',1,1); +INSERT INTO Update_table_3 VALUES (2,'N2',2,2); +INSERT INTO Update_table_3 VALUES (3,'N3',3,3); +INSERT INTO Update_table_3 VALUES (4,'N4',4,4); +INSERT INTO Update_table_3 VALUES (5,'N5',5,5); +INSERT INTO Update_table_3 VALUES (6,'N6',6,6); +INSERT INTO Update_table_3 VALUES (6,'N6',6,6); + +-- echo 1. update-select +UPDATE Update_table_2 SET t_name='N01', col2=2 WHERE id=1; +UPDATE Update_table_2 SET col1=2, col2=(select Update_table_3.col2 from Update_table_3 where Update_table_3.id=1) WHERE id=1; +-- sort SELECT * FROM Update_table_2; + +-- echo 2. update-select rows +UPDATE Update_table_2 SET col1=(select Update_table_3.col2 from Update_table_3) WHERE id=1; +UPDATE Update_table_2 SET col1=(select Update_table_3.col2 from Update_table_3 where Update_table_3.id>1) WHERE id=1; +UPDATE Update_table_2 SET t_name=(select Update_table_3.col2 from Update_table_3 where Update_table_3.id<1) WHERE id=1; +-- sort SELECT * FROM Update_table_2; + +-- echo 3. update rows-select rows +UPDATE Update_table_3 SET col1=(select col2 from Update_table_2); +-- sort SELECT * FROM Update_table_3; + +-- echo 3. update rows-select cols and rows +UPDATE Update_table_3 SET col1=(select col1, col2 from Update_table_2); +-- sort SELECT * FROM Update_table_3; + +-- echo 4. check typecast +UPDATE Update_table_3 SET t_name=813,col1=547.82 WHERE id=2; +UPDATE Update_table_3 SET t_name=(select Update_table_2.col1 from Update_table_2 where Update_table_2.id=3), col2=(select min(Update_table_3.col2) from Update_table_3 where Update_table_3.id=1) where id=5; +-- sort SELECT * FROM Update_table_3; + +-- echo 5. subquery return null +UPDATE Update_table_3 SET col1=2, col2=(select Update_table_2.col2 from Update_table_2 where Update_table_2.id=999) WHERE id=4; +UPDATE Update_table_3 SET col1=2, col2=(select Update_table_2.col2 from Update_table_2 where Update_table_2.id=999) WHERE id=666; +-- sort SELECT * FROM Update_table_3; + +-- echo 6. subquery return multiple rows +UPDATE Update_table_2 SET t_name=(select Update_table_3.t_name from Update_table_3 where Update_table_3.id=6) WHERE col1=5 and col2=3; +UPDATE Update_table_2 SET t_name=(select Update_table_3.t_name from Update_table_3 where Update_table_3.id=6) WHERE col1=5 and col2=6; +-- sort SELECT * FROM Update_table_2; diff --git a/test/case/test/primary-update.test b/test/case/test/primary-update.test index 23405322..4b016ce9 100644 --- a/test/case/test/primary-update.test +++ b/test/case/test/primary-update.test @@ -1,5 +1,5 @@ -- echo initialization -CREATE TABLE Update_table_1(id int, t_name char, col1 int, col2 int); +CREATE TABLE Update_table_1(id int, t_name char(4), col1 int, col2 int); CREATE INDEX index_id on Update_table_1(id); INSERT INTO Update_table_1 VALUES (1,'N1',1,1); INSERT INTO Update_table_1 VALUES (2,'N2',1,1); diff --git a/test/perf/client_performance_test.cpp b/test/perf/client_performance_test.cpp index 25d2ddc7..fa4b8ee0 100644 --- a/test/perf/client_performance_test.cpp +++ b/test/perf/client_performance_test.cpp @@ -29,7 +29,7 @@ See the Mulan PSL v2 for more details. */ #include "common/metrics/metrics.h" #include "common/metrics/metrics_registry.h" -#define MAX_MEM_BUFFER_SIZE 8192 +#define MAX_MEM_BUFFER_SIZE 131072 #define PORT_DEFAULT 6789 using namespace common; diff --git a/unittest/observer/aggregate_hash_table_test.cpp b/unittest/observer/aggregate_hash_table_test.cpp index 6eb62624..c2e997e6 100644 --- a/unittest/observer/aggregate_hash_table_test.cpp +++ b/unittest/observer/aggregate_hash_table_test.cpp @@ -32,7 +32,7 @@ TEST(AggregateHashTableTest, DISABLED_standard_hash_table) group_chunk.add_column(std::move(column1), 0); aggr_chunk.add_column(std::move(column2), 1); - AggregateExpr aggregate_expr(AggregateExpr::Type::SUM, nullptr); + AggregateFunctionExpr aggregate_expr(AggregateFunctionExpr::Type::SUM, nullptr); std::vector aggregate_exprs; aggregate_exprs.push_back(&aggregate_expr); auto standard_hash_table = std::make_unique(aggregate_exprs); @@ -75,7 +75,7 @@ TEST(AggregateHashTableTest, DISABLED_standard_hash_table) aggr_chunk.add_column(std::move(aggr1), 0); aggr_chunk.add_column(std::move(aggr2), 1); - AggregateExpr aggregate_expr(AggregateExpr::Type::SUM, nullptr); + AggregateFunctionExpr aggregate_expr(AggregateFunctionExpr::Type::SUM, nullptr); std::vector aggregate_exprs; aggregate_exprs.push_back(&aggregate_expr); aggregate_exprs.push_back(&aggregate_expr); @@ -120,8 +120,9 @@ TEST(AggregateHashTableTest, DISABLED_linear_probing_hash_table) group_chunk.add_column(std::move(column1), 0); aggr_chunk.add_column(std::move(column2), 1); - auto linear_probing_hash_table = std::make_unique>(AggregateExpr::Type::SUM); - RC rc = linear_probing_hash_table->add_chunk(group_chunk, aggr_chunk); + auto linear_probing_hash_table = + std::make_unique>(AggregateFunctionExpr::Type::SUM); + RC rc = linear_probing_hash_table->add_chunk(group_chunk, aggr_chunk); ASSERT_EQ(rc, RC::SUCCESS); Chunk output_chunk; output_chunk.add_column( @@ -159,7 +160,7 @@ TEST(AggregateHashTableTest, DISABLED_linear_probing_hash_table) aggr_chunk.add_column(std::move(column2), 1); auto linear_probing_hash_table = - std::make_unique>(AggregateExpr::Type::SUM, 256); + std::make_unique>(AggregateFunctionExpr::Type::SUM, 256); RC rc = linear_probing_hash_table->add_chunk(group_chunk, aggr_chunk); ASSERT_EQ(rc, RC::SUCCESS); Chunk output_chunk; diff --git a/unittest/observer/bplus_tree_log_entry_test.cpp b/unittest/observer/bplus_tree_log_entry_test.cpp.disabled similarity index 100% rename from unittest/observer/bplus_tree_log_entry_test.cpp rename to unittest/observer/bplus_tree_log_entry_test.cpp.disabled diff --git a/unittest/observer/bplus_tree_log_test.cpp b/unittest/observer/bplus_tree_log_test.cpp.disabled similarity index 100% rename from unittest/observer/bplus_tree_log_test.cpp rename to unittest/observer/bplus_tree_log_test.cpp.disabled diff --git a/unittest/observer/bplus_tree_test.cpp b/unittest/observer/bplus_tree_test.cpp.disabled similarity index 100% rename from unittest/observer/bplus_tree_test.cpp rename to unittest/observer/bplus_tree_test.cpp.disabled diff --git a/unittest/observer/expression_test.cpp b/unittest/observer/expression_test.cpp index 9b8bb697..e90c3cb7 100644 --- a/unittest/observer/expression_test.cpp +++ b/unittest/observer/expression_test.cpp @@ -281,8 +281,8 @@ TEST(ComparisonExpr, comparison_expr_test) Value int_value(1); FieldMeta field_meta("col1", AttrType::INTS, 0, int_len, true, 0); Field field(nullptr, &field_meta); - unique_ptr right_expr = std::make_unique(field); - int count = 1024; + unique_ptr right_expr = std::make_unique(field); + int count = 1024; std::unique_ptr column_right = std::make_unique(AttrType::INTS, int_len, count); for (int i = 0; i < count; ++i) { int right_value = i; @@ -307,11 +307,11 @@ TEST(ComparisonExpr, comparison_expr_test) } } -TEST(AggregateExpr, aggregate_expr_test) +TEST(AggregateFunctionExpr, aggregate_expr_test) { Value int_value(1); unique_ptr value_expr(new ValueExpr(int_value)); - AggregateExpr aggregate_expr(AggregateExpr::Type::SUM, std::move(value_expr)); + AggregateFunctionExpr aggregate_expr(AggregateFunctionExpr::Type::SUM, std::move(value_expr)); aggregate_expr.equal(aggregate_expr); auto aggregator = aggregate_expr.create_aggregator(); for (int i = 0; i < 100; i++) { @@ -320,18 +320,18 @@ TEST(AggregateExpr, aggregate_expr_test) Value result; aggregator->evaluate(result); ASSERT_EQ(result.get_int(), 4950); - AggregateExpr::Type aggr_type; - ASSERT_EQ(RC::SUCCESS, AggregateExpr::type_from_string("sum", aggr_type)); - ASSERT_EQ(aggr_type, AggregateExpr::Type::SUM); - ASSERT_EQ(RC::SUCCESS, AggregateExpr::type_from_string("count", aggr_type)); - ASSERT_EQ(aggr_type, AggregateExpr::Type::COUNT); - ASSERT_EQ(RC::SUCCESS, AggregateExpr::type_from_string("avg", aggr_type)); - ASSERT_EQ(aggr_type, AggregateExpr::Type::AVG); - ASSERT_EQ(RC::SUCCESS, AggregateExpr::type_from_string("max", aggr_type)); - ASSERT_EQ(aggr_type, AggregateExpr::Type::MAX); - ASSERT_EQ(RC::SUCCESS, AggregateExpr::type_from_string("min", aggr_type)); - ASSERT_EQ(aggr_type, AggregateExpr::Type::MIN); - ASSERT_EQ(RC::INVALID_ARGUMENT, AggregateExpr::type_from_string("invalid type", aggr_type)); + AggregateFunctionExpr::Type aggr_type; + ASSERT_EQ(RC::SUCCESS, AggregateFunctionExpr::type_from_string("sum", aggr_type)); + ASSERT_EQ(aggr_type, AggregateFunctionExpr::Type::SUM); + ASSERT_EQ(RC::SUCCESS, AggregateFunctionExpr::type_from_string("count", aggr_type)); + ASSERT_EQ(aggr_type, AggregateFunctionExpr::Type::COUNT); + ASSERT_EQ(RC::SUCCESS, AggregateFunctionExpr::type_from_string("avg", aggr_type)); + ASSERT_EQ(aggr_type, AggregateFunctionExpr::Type::AVG); + ASSERT_EQ(RC::SUCCESS, AggregateFunctionExpr::type_from_string("max", aggr_type)); + ASSERT_EQ(aggr_type, AggregateFunctionExpr::Type::MAX); + ASSERT_EQ(RC::SUCCESS, AggregateFunctionExpr::type_from_string("min", aggr_type)); + ASSERT_EQ(aggr_type, AggregateFunctionExpr::Type::MIN); + ASSERT_EQ(RC::INVALID_ARGUMENT, AggregateFunctionExpr::type_from_string("invalid type", aggr_type)); } int main(int argc, char **argv) diff --git a/unittest/observer/mvcc_trx_log_test.cpp b/unittest/observer/mvcc_trx_log_test.cpp.disabled similarity index 100% rename from unittest/observer/mvcc_trx_log_test.cpp rename to unittest/observer/mvcc_trx_log_test.cpp.disabled