Summary
This release introduces support for catalogs.yml v2, a skip_optimize model config to bypass post-materialization OPTIMIZE calls, and connection parameter support for the SQL connector's Rust kernel backend. Key fixes resolve issues with dropped constraints, missing column tags and comments on incremental runs and views, and redundant rebuilds of materialized views caused by configuration drift.
Summary generated by brickster.ai. For the full changelog and any code/binary attachments, follow the GitHub link above.
More from databricks/dbt-databricks
1.12.1
This release adds support for Single Point of Gateway hosts and exposes Databricks Jobs runtime IDs in the adapter response to correlate dbt runs with workflow executions. Key fixes resolve issues with capability-branching macros, column-level tags on snapshots, and liquid clustering on managed Iceberg models, while a breaking change now requires setting contract enforcement to true for column-level constraints to apply.
This release introduces support for metric views, row filters, Python UDFs, key-only table and column tagging, and new refresh modes for materialized views and streaming tables. Additionally, databricks_tags now merge additively across hierarchy levels instead of child configurations completely replacing parent configurations.
This release adds the invocation ID to the default query comment and enforces a 255-character identifier length limit for Databricks relations. It also resolves several issues, including fixing materialized view refreshes when tags are set, warning on enforced contracts for materialized views, and suppressing several spurious warnings.
This release introduces notebook-scoped packages for command submits and notebook job runs, alongside new insert-by-name support for microbatch and replace_where strategies. Key fixes resolve duplicate aliases in --empty mode, correct workflow job spec conversions, and improve capability detection for named compute.
This release enables concurrent microbatch execution and adds optimize calls to both snapshot and table v2 materializations. Key fixes include quoting catalog names in SHOW SCHEMAS commands, adding cluster clauses to streaming table alter SQL, and applying column-level tags for V1 table materialization.