Repository navigation

7.0: parallel XLSX/XLSB reading, Excel tables, own byte sources, managed inflate, and a public API cleanup - #140

Merged

GabrielMarquezMatte merged 121 commits into

masterfrom

develop

Oct 12, 2026

Merged

7.0: parallel XLSX/XLSB reading, Excel tables, own byte sources, managed inflate, and a public API cleanup#140
GabrielMarquezMatte merged 121 commits into
masterfrom
develop

Commits on Oct 7, 2026
- chore: promote public API for v6.0.0
  
  GabrielMarquezMatte
  authored and
  github-actions[bot]
  committed
  6fcd15c
- Merge pull request #137 from GabrielMarquezMatte/chore/promote-public-api-v6.0.0
  GabrielMarquezMatte
  authored
  d3a3c3a
- feat(python): enhance Arrow stream handling with GIL management and threading support
  GabrielMarquezMatte
  committed
  d86e4af
- Merge branch 'develop' of https://github.com/GabrielMarquezMatte/ExcelReader into develop
  GabrielMarquezMatte
  committed
  5947e6e
- feat(core): open a workbook from any byte source through a seekable view
  
  GabrielMarquezMatte
  and
  claude
  committed
  57dc08c
- feat(core): cache a byte source in blocks fetched once
  
  GabrielMarquezMatte
  and
  claude
  committed
  fcb2605
- feat(native): add source and buffering options and keep accepting the 6.0 options layout
  
  GabrielMarquezMatte
  and
  claude
  committed
  5574daa
- fix(native): make the 6.0 options size a constant and document it
  
  GabrielMarquezMatte
  and
  claude
  committed
  d46ec61
- feat(native): open a workbook over a source the caller serves
  
  GabrielMarquezMatte
  and
  claude
  committed
  6bc46a0
- feat(native): open a workbook over a stream that cannot seek
  
  GabrielMarquezMatte
  and
  claude
  committed
  0e20232
- test(native): open a workbook through xl_open_source from C
  
  GabrielMarquezMatte
  and
  claude
  committed
  4714b45
- feat(rust): open a workbook from a Source or any Read
  
  GabrielMarquezMatte
  and
  claude
  committed
  c933af8
- test(rust): pin source release counts and the new ABI layouts
  
  GabrielMarquezMatte
  and
  claude
  committed
  962e872
- feat(cpp): open a workbook from an xl::Source or an xl::InputStream
  
  GabrielMarquezMatte
  and
  claude
  committed
  4db2192
- feat(python): open a workbook from a Source or a readable stream
  
  GabrielMarquezMatte
  and
  claude
  committed
  2b1b3ad
- fix(python): register a source only after every fallible step, and guard stream reads
  
  GabrielMarquezMatte
  and
  claude
  committed
  995dec9
- docs: read workbooks from your own sources and streams
  
  GabrielMarquezMatte
  and
  claude
  committed
  fe9a3b6
- fix: final-review fixes for callback sources
  
  GabrielMarquezMatte
  and
  claude
  committed
  79f6201
- feat: enhance CSV stream handling to support schema inference and replay functionality
  GabrielMarquezMatte
  committed
  e269dc6
- Bump the analyzers group with 1 update
  dependabot[bot]
  authored
  d5da1e1
- Bump the test group with 2 updates
  dependabot[bot]
  authored
  d185665
- Removing unnecessary comments
  GabrielMarquezMatte
  committed
  77f97b6
Commits on Oct 8, 2026
- Merge pull request #138 from GabrielMarquezMatte/dependabot/nuget/develop/analyzers-55e46a7a66
  GabrielMarquezMatte
  authored
  4f8535d
- Merge pull request #139 from GabrielMarquezMatte/dependabot/nuget/tests/ExcelReader.Tests/develop/test-4352e14058
  GabrielMarquezMatte
  authored
  1975041
- feat: add Huffman decode-table builder for the managed inflate decoder
  
  GabrielMarquezMatte
  and
  claude
  committed
  1514726
- feat: add a managed streaming inflate decoder
  
  GabrielMarquezMatte
  and
  claude
  committed
  d7f1559
- perf: inflate ZIP entries with the managed decoder
  
  GabrielMarquezMatte
  and
  claude
  committed
  7f02562
- test: fuzz the managed inflate decoder against DeflateStream
  
  GabrielMarquezMatte
  and
  claude
  committed
  58bb3ca
- perf: benchmark the managed inflate decoder against DeflateStream
  
  GabrielMarquezMatte
  and
  claude
  committed
  0929fc6
- docs: record the managed inflate measurements
  
  GabrielMarquezMatte
  and
  claude
  committed
  408b76e
- fix: reject a truncated deflate stream before handing out padding-decoded bytes
  
  GabrielMarquezMatte
  and
  claude
  committed
  2ad7123
- fix(fuzz): let inflate divergences and canary trips escape the oracle
  
  GabrielMarquezMatte
  and
  claude
  committed
  0579fa4
- fix: update binary file for ExcelReader native integration
  GabrielMarquezMatte
  committed
  5997074
- perf: tighten the inflate fast loop
  
  GabrielMarquezMatte
  and
  claude
  committed
  ab3aa1e
- perf(native): regenerate the static PGO profile after the inflate fast-loop change
  
  GabrielMarquezMatte
  and
  claude
  committed
  1d3afab
- Merge branch 'develop' of https://github.com/GabrielMarquezMatte/ExcelReader into develop
  GabrielMarquezMatte
  committed
  640a6e2
- perf(native): build typed columns in native memory and hand them over without a copy
  
  GabrielMarquezMatte
  and
  claude
  committed
  782ad91
- Merge branch 'master' into develop
  GabrielMarquezMatte
  committed
  fc6e38a
- perf(native): regenerate the static PGO profile after the native column buffers
  
  GabrielMarquezMatte
  and
  claude
  committed
  d2ec3cf
- ci: run InflateBenchmark with the read group
  
  GabrielMarquezMatte
  and
  claude
  committed
  df3634c
- fix(generator): set required members in the generated factory
  
  GabrielMarquezMatte
  and
  claude
  committed
  7a08ec2
Commits on Oct 9, 2026
- feat(native): add support for auto-detecting CSV format and implement benchmarks for native open operations
  GabrielMarquezMatte
  committed
  65945b4
- perf(docs): update benchmark results for improved performance metrics across C++, Python, and Rust implementations
  GabrielMarquezMatte
  committed
  415d97f
- perf(docs): update benchmark results for XLSX, XLSM, XLSB, and CSV formats to reflect improved performance metrics
  GabrielMarquezMatte
  committed
  7e04748
- perf(docs): update performance metrics for NativeAOT XLSX and CSV enumerator overhead based on recent measurements
  GabrielMarquezMatte
  committed
  4aa3766
- feat(reader): add table part model with A1 range parsing
  
  GabrielMarquezMatte
  and
  claude
  committed
  f90a080
- feat(reader): expose workbook tables on every format
  
  GabrielMarquezMatte
  and
  claude
  committed
  c402ebc
- feat(reader): discover XLSX tables at open
  
  GabrielMarquezMatte
  and
  claude
  committed
  d60068b
- feat(reader): discover XLSB tables at open
  
  GabrielMarquezMatte
  and
  claude
  committed
  48e0aed
- fix(reader): discover XLSB tables before renting shared strings
  
  GabrielMarquezMatte
  and
  claude
  committed
  bdf0e2f
- feat(reader): track absolute row index on demand in XLSX/XLSB enumerators
  
  GabrielMarquezMatte
  and
  claude
  committed
  ed16d72
Commits on Oct 10, 2026
- feat(reader): read a table's rows through ExcelTable.AsSheet
  
  GabrielMarquezMatte
  and
  claude
  committed
  d2b30d7
- feat(reader): reject name binding on headerless tables
  
  GabrielMarquezMatte
  and
  claude
  committed
  a24214f
- test(fuzz): drain table views; docs: document reading Excel tables
  
  GabrielMarquezMatte
  and
  claude
  committed
  fac7938
- fix(reader): map table view columns from the table's first column
  
  GabrielMarquezMatte
  and
  claude
  committed
  1393c4c
- fix(reader): reject malformed row numbers in table views
  
  GabrielMarquezMatte
  and
  claude
  committed
  325784b
- docs: clarify headerless-table and column-index rules for table views
  
  GabrielMarquezMatte
  and
  claude
  committed
  addef34
- fix(reader): keep XLSX rows whose open tag straddles an async refill
  
  GabrielMarquezMatte
  and
  claude
  committed
  cf27e69
- feat(writer): add table options, validation and tracking
  
  GabrielMarquezMatte
  and
  claude
  committed
  4a4519c
- feat(writer): write Excel tables to XLSX
  
  GabrielMarquezMatte
  and
  claude
  committed
  8512d9e
- test(writer): cover the async end path for XLSX tables
  
  GabrielMarquezMatte
  and
  claude
  committed
  528c573
- feat(writer): write Excel tables to XLSB
  
  GabrielMarquezMatte
  and
  claude
  committed
  5842fa7
- feat(writer): write records as an Excel table; docs: document writing tables
  
  GabrielMarquezMatte
  and
  claude
  committed
  975d55d
- fix(writer): drop the stray sheet AutoFilter and sort state from XLSB sheets
  
  GabrielMarquezMatte
  and
  claude
  committed
  155f25b
- fix(writer): keep the reserved row of a header-only table out of the next table
  
  GabrielMarquezMatte
  and
  claude
  committed
  7634308
- fix(writer): escape XLSX table column names like their header cells
  
  GabrielMarquezMatte
  and
  claude
  committed
  470f8c2
- docs: PrefetchDecompression applies to in-memory XLSX/XLSB reads
  
  GabrielMarquezMatte
  and
  claude
  committed
  2fd4f7a
- chore: add the project icon to the NuGet packages and the README
  GabrielMarquezMatte
  committed
  fa71c84
- chore: show the project icon in the Python, Rust and C++ READMEs and on docs.rs
  GabrielMarquezMatte
  committed
  30edcd0
Commits on Oct 11, 2026
- refactor!: rename the parallel record, accumulator and aggregation contracts off their CSV names
  
  GabrielMarquezMatte
  and
  claude
  committed
  7e0a8db
- feat: add SheetParallel, a parallel reading API that takes a sheet of any format
  
  GabrielMarquezMatte
  and
  claude
  committed
  97d797c
- feat!: remove CsvParallel in favour of SheetParallel
  
  GabrielMarquezMatte
  and
  claude
  committed
  cd6550e
- feat(reader): let XLSX and XLSB enumerators parse a slice of inflated sheet bytes in place
  
  GabrielMarquezMatte
  and
  claude
  committed
  6b49056
- feat: cut an inflated sheet stream into row-aligned chunks
  
  GabrielMarquezMatte
  and
  claude
  committed
  ddac3c8
- fix: stop the XLSX chunk producer at the end of sheet data
  
  GabrielMarquezMatte
  and
  claude
  committed
  c2ec1c1
- feat: add the sheet pipeline, reading XLSX rows on several threads
  GabrielMarquezMatte
  committed
  8aa85e8
- feat: read XLSX sheets in parallel through SheetParallel
  
  GabrielMarquezMatte
  and
  claude
  committed
  5b3d755
- fix: report the sheet row for a parse failure after pipeline recovery
  
  GabrielMarquezMatte
  and
  claude
  committed
  6749b9e
- feat: read XLSB sheets in parallel through SheetParallel
  
  GabrielMarquezMatte
  and
  claude
  committed
  04e084c
- feat(native): parse XLSX and XLSB typed tables in parallel
  
  GabrielMarquezMatte
  and
  claude
  committed
  01369bd
- test(fuzz): compare the sheet pipeline with the sequential read
  
  GabrielMarquezMatte
  and
  claude
  committed
  0696337
- perf: cut sheet pipeline chunks at 256 KB instead of 1 MB
  
  GabrielMarquezMatte
  and
  claude
  committed
  d6b1186
- perf: benchmark parallel sheet reading and regenerate the static PGO profile
  
  GabrielMarquezMatte
  and
  claude
  committed
  fa57c54
- docs: mark the parallel benchmark tables as pending an idle rerun
  
  GabrielMarquezMatte
  and
  claude
  committed
  580fba6
- docs: document SheetParallel and the v7 migration
  
  GabrielMarquezMatte
  and
  claude
  committed
  c9c3dc6
- fix: surface a source stream's own cancellation from the sheet pipeline
  
  GabrielMarquezMatte
  and
  claude
  committed
  0ffacde
- docs: correct the CSV fallback rule and sample variables
  
  GabrielMarquezMatte
  and
  claude
  committed
  876d9ac
- fix(reader): decode OOXML _xHHHH_ escapes in XLSX text
  
  GabrielMarquezMatte
  and
  claude
  committed
  e56971d
- fix(reader): reject a streamed ZIP entry that ends short of its declared size
  
  GabrielMarquezMatte
  and
  claude
  committed
  60686eb
- refactor!: rename ExcelParserConfig to ExcelParserOptions
  GabrielMarquezMatte
  committed
  11eb8c8
- refactor!: hand converters an ExcelRowContext instead of loose date-system and culture arguments
  GabrielMarquezMatte
  committed
  0ece1fe
- refactor!: take options objects instead of positional optional parameters
  GabrielMarquezMatte
  committed
  042c8a3
- refactor!: give the XLS writer an options record and the XLSX writer a 1904 date system
  GabrielMarquezMatte
  committed
  30af28e
- refactor!: open CSV with ExcelReaderOptions like every other format
  GabrielMarquezMatte
  committed
  fda69b8
- docs: list the v7 options and converter-context changes in the migration guide
  GabrielMarquezMatte
  committed
  3f9fff8
- refactor!: return the inferred schema as IReadOnlyList<ExcelColumnSchema>
  GabrielMarquezMatte
  committed
  1cc50ae
- refactor!: remove Cell.TryGetDateTime(out DateTime), which assumed the 1900 date system
  GabrielMarquezMatte
  committed
  b2891ff
- refactor!: rename ExcelDataReader to SheetDataReader
  GabrielMarquezMatte
  committed
  b1f0980
- refactor!: type formula cells by their cached result and remove CellType.Formula
  GabrielMarquezMatte
  committed
  584acd4
- test(python): tie the GIL test's duration floor to the switch interval
  GabrielMarquezMatte
  committed
  4569c30
- perf: load the next litlen entry before the refill that follows a match
  GabrielMarquezMatte
  committed
  fff70b3
- perf(native): regenerate the static PGO profile after the inflate fast-loop change
  GabrielMarquezMatte
  committed
  caa4038
- refactor: optimize XML parsing logic in XlsxWorkbook and XlsxXml classes
  GabrielMarquezMatte
  committed
  c053bfd
Commits on Oct 12, 2026
- fix(rust): let docs.rs build the crate without the native library
  
  GabrielMarquezMatte
  and
  claude
  committed
  f3e104a
- docs(rust): use the README as the crate documentation and compile its examples
  
  GabrielMarquezMatte
  and
  claude
  committed
  79f7fda
- build(generator): stop adding the analyzer release files twice
  
  GabrielMarquezMatte
  and
  claude
  committed
  0a02d7e
- docs(core): inherit only the summary and exceptions in BeginTableAsync
  
  GabrielMarquezMatte
  and
  claude
  committed
  7545986
- docs: fix the AOT anchor in the parsing guide
  
  GabrielMarquezMatte
  and
  claude
  committed
  4e3d978
- docs: publish a docfx site with the API reference and guides
  
  GabrielMarquezMatte
  and
  claude
  committed
  2d31fa7
- docs: add a guide mapping errors across languages
  
  GabrielMarquezMatte
  and
  claude
  committed
  128fa43
- docs(python): generate the Python API reference for the docfx site
  
  GabrielMarquezMatte
  and
  claude
  committed
  1fb0ad1
- docs(cpp): generate the C++ API reference from Doxygen XML
  
  GabrielMarquezMatte
  and
  claude
  committed
  30f640d
- docs(cpp): document the public C++ API with Doxygen comments
  
  GabrielMarquezMatte
  and
  claude
  committed
  776d352
- ci(docs): fail on Doxygen warnings and build the docs on pull requests
  
  GabrielMarquezMatte
  and
  claude
  committed
  0911fdb
- Enhance documentation with cross-references and clarifications
  GabrielMarquezMatte
  committed
  58a279b
- Add snippets for CSV reading, writing, and encryption examples
  GabrielMarquezMatte
  committed
  c957491
- ci(docs): build generator project before site generation to resolve analyzer issues
  GabrielMarquezMatte
  committed
  0fde713
- docs: add guidance for reading untrusted files and implement related snippets
  GabrielMarquezMatte
  committed
  53e31e8
- docs: clarify limits for decompressed bytes in untrusted file handling and update related code comments
  GabrielMarquezMatte
  committed
  405e195
- ci: optimize workflow triggers by limiting branches for benchmarks and native bindings
  GabrielMarquezMatte
  committed
  518c2d7
- ci: disable analyzers during builds to improve performance across workflows
  GabrielMarquezMatte
  committed
  b796cfc

Commits

Commits on Oct 7, 2026

Commits on Oct 8, 2026

Commits on Oct 9, 2026

Commits on Oct 10, 2026

Commits on Oct 11, 2026

Commits on Oct 12, 2026