Overview

The DuckLake project is governed by the non-profit DuckDB Foundation.

Planned Features (Last Updated: September 2025)

This section lists the features that we plan to work on this year. The list was compiled by the DuckLake maintainers and is based on the long-term vision for the project and general interactions with users in the open-source community.

• Documentation regarding access control in DuckLake. ✅
• Migration guides for moving from DuckDB to DuckLake, covering both DuckDB and PostgreSQL catalogs. ✅
• Optimize heavily deleted tables. ✅
• Cleanup of orphan files. ✅
• Geometry/geospatial types. ✅
• Fine-grained update conflicts at the file level to avoid many retries when multiple writers are updating the same table.
• Data inlining with PostgreSQL as a catalog.
• Scalar and table macros.
• Read performance improvements regarding parquet bloom filters and metadata scans. See 1, 2.

Please note that there are no guarantees that a particular feature will be released within the next release. Everything on this page is subject to change without notice.

Future Work

These are some of the items that we plan to support in the future:

• Allow primary key syntax without really enforcing it, similar to what other OLAP engines do (e.g., BigQuery).
• User defined types.
• Variant types.
• Fixed-size arrays.
• Default values that are not literals.