duckdb/ducklake

Health v2 · maintenance only

Flags: no_releases no_readme

Adoption not part of the score

2939 stars · 240 forks observed · 2026-08-28

What it is AI-extracted, prompt v1, taxonomy v1, 2026-08-30, confidence not recorded

DuckLake is an integrated data lake and lakehouse catalog format that stores data as Parquet files and manages metadata in an ACID-compliant SQL database such as DuckDB, SQLite, or PostgreSQL. It ships as a DuckDB extension providing snapshots, time travel, schema evolution, partitioning, and multi-client concurrent access.

Use cases

• build a lakehouse with a SQL-based catalog
• run time travel queries on a data lake
• share a dataset across multiple DuckDB instances with concurrent writes
• replace a single large DuckDB file with partitioned Parquet storage
• migrate from DuckDB files or Iceberg to a simpler lakehouse format
• store data lake files in object storage with ACID guarantees

When to choose

• you want lakehouse features like snapshots, schema evolution, and ACID transactions without Iceberg/Delta complexity
• you already use DuckDB and need multiplayer concurrent access to shared data
• you prefer a SQL database as the catalog with no vendor lock-in

When to avoid

• you need a catalog server independent of SQL databases or non-SQL catalog ecosystems
• your stack is Spark-centric with deep Iceberg/Delta Lake tooling integrations
• you only need a single-user embedded database with no lake requirements

Facets

library · maturity active

database etl analytics databases big-data analytics cpp python cli cross-platform self-hosted lakehouse-format open-table-format parquet duckdb-extension data-lake acid-transactions time-travel schema-evolution catalog data-engineering

Member repositories

For agents

markdown · JSON · MCP: product_card(name="duckdb/ducklake")