ivy_mooncake 🥮

Real-time analytics on Postgres tables (IvorySQL distribution)

Overview

ivy_mooncake is the IvorySQL distribution of pg_mooncake, a Postgres extension that creates a columnstore mirror of your Postgres tables in Iceberg, enabling fast analytics queries with sub-second freshness:

• Real-time ingestion powered by moonlink for streaming and batched INSERT/UPDATE/DELETE.
• Fast analytics accelerated by DuckDB, ranking top 10 on ClickBench.
• Postgres-native allowing you to query a columnstore table just like a regular Postgres table.
• Iceberg-native making your data readily accessible by other query engines.

This fork tracks IvorySQL-maintained branches of pg_duckdb, moonlink, and duckdb_mooncake (ivy_duckdb, ivy_moonlink, ivy_duckdb_mooncake). The repository name is ivy_mooncake, but the extension it installs is unchanged — pg_mooncake — so SQL-side compatibility with upstream is preserved.

Installation

To build ivy_mooncake, first install Rust, pgrx, and the build tools for DuckDB.

Then, clone the repository with submodules:

git clone --recurse-submodules https://github.com/IvorySQL/ivy_mooncake.git

To build and install for Postgres versions 14-18, run:

cargo pgrx init --pg18=$(which pg_config)   # Replace with your Postgres version
make ivy_duckdb                             # Skip if ivy_duckdb is already installed
make install PG_VERSION=pg18

Finally, add pg_mooncake to shared_preload_libraries in your postgresql.conf file and enable logical replication:

duckdb.allow_community_extensions = true
shared_preload_libraries = 'pg_duckdb,pg_mooncake'
wal_level = logical

Quick Start

First, create the pg_mooncake extension:

CREATE EXTENSION pg_mooncake CASCADE;

Next, create a regular Postgres table trades:

CREATE TABLE trades(
  id bigint PRIMARY KEY,
  symbol text,
  time timestamp,
  price real
);

Then, create a columnstore mirror trades_iceberg that stays in sync with trades:

CALL mooncake.create_table('trades_iceberg', 'trades');

Now, insert some data into trades:

INSERT INTO trades VALUES
  (1,  'AMD', '2024-06-05 10:00:00', 119),
  (2, 'AMZN', '2024-06-05 10:05:00', 207),
  (3, 'AAPL', '2024-06-05 10:10:00', 203),
  (4, 'AMZN', '2024-06-05 10:15:00', 210);

Finally, query trades_iceberg to see that it reflects the up-to-date state of trades:

SELECT avg(price) FROM trades_iceberg WHERE symbol = 'AMZN';

Note: The repository is renamed to ivy_mooncake only at the GitHub-fork level. Inside the repo, the extension, the cdylib, and all SQL-level identifiers (pg_mooncake extension, mooncake.* schema, USING mooncake access method, INSTALL mooncake FROM community) are intentionally unchanged from upstream so SQL written against pg_mooncake continues to work.

Upstream

ivy_mooncake is a fork of Mooncake-Labs/pg_mooncake under the MIT License.