github databricks/dbt-databricks v1.13.1

one hour ago

Features

  • Accept use_materialization_v2 as a model config, so a project or an individual table or incremental model can choose between the atomic CREATE OR REPLACE TABLE … AS SELECT path and the create-then-insert path; models that don't set it follow the use_materialization_v2 behavior flag, and view models ignore it by @sd-db in #1718

Fixes

  • Stop dbt docs generate from emitting a RuntimeWarning per schema for catalog metadata columns that are absent from the results (thanks @emrekabalcii!) (#1749 resolves #1748)
  • Unset tags on columns that a table model or incremental full refresh removes, so replacing a relation no longer fails with CANNOT_DROP_TAGGED_COLUMN when a removed column carries a governed tag by @sd-db in #1707 (resolves #1684)
  • Load wide seeds faster by resolving the placeholder and parquet cast types once per seed instead of once per cell by @sd-db in #1722
  • Stop materialized views with a Compatibility Mode location from being dropped and lost when use_safer_relation_operations or a partition_by change would recreate them at the same location. Unchanged partitions now use CREATE OR REPLACE, and location or partition changes fail before anything is dropped by @sd-db in #1727
  • Make the new table visible to the model's post-hooks and the rest of the run when a table or incremental model replaces a view, materialized view, or streaming table by @sd-db in #1708
  • Stop dbt clone from failing on Unity Catalog tables with a row filter or column mask, which can't be shallow-cloned; they now clone as a view over the deferred table that applies its policies for whoever queries it by @sd-db in #1745
  • Stop dbt clone from failing on materialized views with a row filter and streaming tables with a row filter or column mask by @sd-db in #1744
  • Stop dbt clone from failing on views whose enforced contract changed since the deferred state by @sd-db in #1742 (resolves #1349)
  • Make is_string(), is_number(), is_integer(), is_float(), and is_numeric() recognize Databricks column types such as string, int, tinyint, and decimal(p,s) by @sd-db in #1726 (resolves #1380)
  • Hide credential-like keys nested inside session_properties from dbt debug connection output by @sd-db in #1728 (resolves #1724)
  • Fix unit tests and direct get_columns_in_query calls failing with PARSE_SYNTAX_ERROR when a model's SQL ends with a semicolon by @sd-db in #1729 (resolves #1725)

Under the Hood

  • Allow dbt-core 1.12.7 by @sd-db in #1776
  • Allow databricks-sql-connector 4.9.x (upper bound raised to <4.9.1) and keep the enable_dbt_telemetry server-flag check working with connector 4.7.0+ by @sd-db in #1768
  • Raise the databricks-sdk upper bound to <0.158.0 by @sd-db in #1770
  • Allow packaging 26.x (upper bound raised to <27.0) by @sd-db in #1777

New Contributors

Full Changelog: v1.13.0...v1.13.1

Don't miss a new dbt-databricks release

NewReleases is sending notifications on new releases.