Pg lake

lastmod 2025-11-06created 2025-11-061 edits

Extension that brings Iceberg and data lake access to Postgres

About

• pg_lake is a Postgres-based extension that directly integrates Iceberg tables and data lake files, supporting transactions and fast queries
• Can directly query, import, and export Parquet, CSV, JSON, and Iceberg files in object storage such as S3
• Uses the DuckDB query engine internally for fast execution within the Postgres environment
• Provides data lakehouse features through a single SQL interface, such as creating Iceberg tables, automatic schema inference for external files, and S3 input/output via the COPY command
• Open-sourced by Snowflake after acquiring Crunchy Data in 2025; a foundation for extending data lake integration in the Postgres ecosystem

See also

• Postgres
• Iceberg
• Data Lake

Favorite site

• Github - Snowflake-Labs/pg_lake
• Pg_lake - Extension that brings Iceberg and data lake access to Postgres | GeekNews