v1.12.3

dbt-databricks 1.12.3 (Jul 29, 2026)

Features

• Support catalog_database in v2 catalogs.yml to route Unity catalog models to a physical catalog independent of the dbt catalog name (requires dbt-core>=1.12 and dbt-adapters>=1.24.4). (thanks @aahel!) (#1590)

Fixes

• Prevent concurrent connection opens from racing during lazy SDK configuration initialization by @sd-db in #1606
• Avoid treating unchanged Streaming Table databricks_tags as configuration changes by diffing against existing table tags. by @sd-db in #1602 (resolves #1601)
• Allow dropping a column that has governed tags by @saishreeeee in #1597 (resolves #1323)
• Fix view materialization incorrectly producing a no-op instead of forcing recreation when --full-refresh is provided alongside view_update_via_alter: true and use_materialization_v2: true (thanks @aarushisingh04!) (#1456 resolves #1404)
• Support dbt clone and rebuilds over an existing shallow clone by @saishreeeee in #1592 (resolves #1165, follow-up by @sd-db in #1608)
• Fix managed Iceberg Python models failing with MANAGED_TABLE_FORMAT by emitting .format("iceberg") instead of the parquet sentinel from resolve_file_format (thanks @Divya-Kovvuru-0802!) (#1593 resolves #1591)
• Quote generated column identifiers in incremental strategies so non-ASCII column names no longer fail on subsequent runs (thanks @ash2shukla!) (#1595 resolves #1594)
• Handle missing or empty view-definition metadata when creating materialized views from streaming tables or newly-created materialized views (thanks @aarushisingh04!) (#1462 resolves #1459)

Under the Hood

• Remove the unused keyring dependency and its transitive packages by @sd-db in #1588
• Raise the dbt-core upper bound to <1.12.1 to include dbt-core 1.12.0 by @sd-db in #1605

New Contributors

• @saishreeeee made their first contribution in #1587
• @ash2shukla made their first contribution in #1595
• @Divya-Kovvuru-0802 made their first contribution in #1593

Full Changelog: v1.12.2...v1.12.3