databricks-iceberg

You can expose Databricks managed tables to Snowflake by creating managed Iceberg tables or enabling UniForm for Delta tables, then configuring external engine access through the Unity Catalog Iceberg REST Catalog (IRC) endpoint with proper credential vending and grants.

Yes, external engines like PyIceberg, OSS Spark, and Trino can read Databricks-managed Iceberg data by connecting to the Unity Catalog Iceberg REST Catalog endpoint, requiring correct network allowlists and authentication via PAT or OAuth.

Unity Catalog supports cross-engine interoperability via the Iceberg REST Catalog endpoint located at /api/2.1/unity-catalog/iceberg-rest, enabling external engines to read and write Databricks-managed Iceberg data with credential vending.

Managed Iceberg tables on Databricks do not support expression transforms in PARTITIONED BY clauses, and you must correctly handle deletion vectors and row tracking for Iceberg v2 CLUSTER BY operations.

You can enable external Iceberg reads for existing Delta streaming tables and materialized views by using Compatibility Mode alongside UniForm, allowing external engines to access Delta data as Iceberg.

Yes, a Unity Catalog-enabled workspace is required to create and manage Apache Iceberg tables on Databricks and expose them to external engines via the Iceberg REST Catalog endpoint.