GitHub - duckdb/ducklake: DuckLake is an integrated data lake and catalog format

DuckDB DuckLake Extension DuckLake is an open Lakehouse format that is built on SQL and Parquet. DuckLake stores metadata in a catalog database, and stores data in Parquet files. The DuckLake extension allows DuckDB to directly read and write data from DuckLake. See the DuckLake website for more information. Installation DuckLake can be installed using the INSTALL command: INSTALL ducklake; The latest development version can be installed from core_nightly: FORCE INSTALL ducklake FROM core_nightly; Usage DuckLake databases can be attached using the ATTACH syntax, after which tables can be created, modified and queried using standard SQL. Below is a short usage example that stores the metadata in a DuckDB database file called metadata.ducklake, and the data in Parquet files in the file_path directory: ATTACH 'ducklake:metadata.ducklake' AS my_ducklake (DATA_PATH 'file_path/'); USE my_ducklake; CREATE TABLE my_ducklake.my_table(id INTEGER, val VARCHAR); INSERT INTO my_ducklake.my_table VALUES (1, 'Hello'), (2, 'World'); FROM my_ducklake.my_table; ┌───────┬─────────┐ │ id │ val │ │ int32 │ varchar │ ├───────┼─────────┤ │ 1 │ Hello │ │ 2 │ World │ └───────┴─────────┘ Updates UPDATE my_ducklake.my_table SET val='DuckLake' WHERE id=2; FROM my_ducklake.my_table; ┌───────┬──────────┐ │ id │ val │ │ int32 │ varchar │ ├───────┼──────────┤ │ 1 │ Hello │ │ 2 │ DuckLake │ └───────┴──────────┘ Time Travel FROM my_ducklake.my_table AT (VERSION => 2); ┌───────┬─────────┐ │ id │ val │ │ int32 │ varchar │ ├───────┼─────────┤ │ 1 │ Hello │ │ 2 │ World │ └───────┴─────────┘ Schema Evolution ALTER TABLE my_ducklake.my_table ADD COLUMN new_column VARCHAR; FROM my_ducklake.my_table; ┌───────┬──────────┬────────────┐ │ id │ val │ new_column │ │ int32 │ varchar │ varchar │ ├───────┼──────────┼────────────┤ │ 1 │ Hello │ NULL │ │ 2 │ DuckLake │ NULL │ └───────┴──────────┴────────────┘ Change Data Feed FROM my_ducklake.table_changes('my_table', 2, 2); ┌─────────────┬───────┬─────────────┬───────┬─────────┐ │ snapshot_id │ rowid │ change_type │ id │ val │ │ int64 │ int64 │ varchar │ int32 │ varchar │ ├─────────────┼───────┼─────────────┼───────┼─────────┤ │ 2 │ 0 │ insert │ 1 │ Hello │ │ 2 │ 1 │ insert │ 2 │ World │ └─────────────┴───────┴─────────────┴───────┴─────────┘ See the Usage guide for more information. Building & Loading the Extension To build, type git submodule init git submodule update --recursive # to build with multiple cores, use `make GEN=ninja release` make The submodules are pinned to the DuckDB version in .github/duckdb-version, which is what CI builds against. make pull moves them to the tip of their branch instead, where the build can fail. To run, run the bundled duckdb shell: ./build/release/duckdb Contributing We are happy to welcome external contributions to our DuckLake extension. Our active development branch is currently main, and all contributions should target that branch. Testing # run all DuckLake extension tests ./build/release/test/unittest # run a single test file ./build/release/test/unittest test/sql/transaction/create_conflict.test # run tests matching a pattern ./build/release/test/unittest "test/sql/partitioning/*" # run DuckDB core tests using DuckLake as storage backend ./build/release/test/unittest --test-config test/configs/attach_ducklake.json --test-dir duckdb # run DuckLake tests using PostgreSQL as catalog database (requires running PostgreSQL) ./build/release/test/unittest --test-config test/configs/postgres.json # run DuckLake tests using SQLite as catalog database ./build/release/test/unittest --test-config test/configs/sqlite.json # run tests with deletion vectors enabled ./build/release/test/unittest --test-config test/configs/deletion_vectors.json