Feature Engineering in Unity Catalog

Feature Engineering in Unity Catalog is a Databricks ai / ml capability, introduced 2023.

Workspace Feature Store → Feature Engineering in Unity Catalog

Feature management built on Unity Catalog - any Delta table with a primary key can serve as a feature table, with UC lineage and governance.

• A feature table is nothing more than a Delta table with a primary key constraint in Unity Catalog - updated, browsed, and governed through Unity Catalog like any other table. 🔗
• Its Python package is databricks-feature-engineering, renamed from the older databricks-feature-store. 🔗

⚠ Limitations: A FeatureSpec can contain at most 1000 features, at most 50 tables can train a model, and at most 100 on-demand features can be used in a model; Databricks Runtime ML clusters can't be used with Lakeflow pipelines (standard access mode clusters are required). 🔗

Category

AI / ML

Introduced

2023🔗

Also known as

Feature Engineering, databricks-feature-engineering

Verified

2026-07-23