Skip to content

v1.4.0

Pre-release
Pre-release
Compare
Choose a tag to compare
@tkirschke tkirschke released this 26 Aug 12:17
· 133 commits to main since this release
33d1a3c

Important Changes

What's new?

All adapters

  • Global variables for rsrc_default_dtype, timestamp_default_dtype, stg_default_dtype, and derived_column_default_dtype are now adapter specific
  • Adapter specific column escaping
  • Enforcing default datatype of ldts: The load date timestamp column is explicitly cast to the specified timestamp_default_dtype

PostgreSQL

  • Usage of adapter specific end_of_all_times

What's changed?

Snowflake

  • Huge improvements in column name escaping, to support all kinds of column names

PostgreSQL

  • Default Hash Datatype set to VARCHAR(32) instead of STRING
  • Broader support of Datatypes in Ghost Record generation
  • Improvement of incremental loading for NH_Links
  • Fix hash macro parameters #207
  • Fix ref_sat_v0 and sat_v0 macro parameters #206
  • Fix control_snap_v0 macro parameters #205

Redshift

  • Default Hash Datatype set to VARCHAR(32) instead of STRING
  • Broader support of Datatypes in Ghost Record generation #204
  • Improvement of incremental loading for NH_Links
  • Usage of DATEADD() instead of TIMESTAMPSUB() in PIT #155
  • CROSS JOIN instead of FULL OUTER JOIN in ref_table
  • Fix hash macro parameters #207
  • Fix ref_sat_v0 and sat_v0 macro parameters #206
  • Fix control_snap_v0 macro parameters #205

Synapse

  • Align the snapshot control tables v0 and v1 with the other adapters

Bugfixes

All adapters

  • Usage of dbt.type_timestamp() instead of just type_timestamp()

Postgres

  • Timestamp format for Postgres changed to YYYY-MM-DD HH24:MI:SS
  • Fixed Interval issues in Control_Snap_v1
  • Added option for disable_hwm and source_is_single_batch to sat_v0 to align with other adapters

Redshift:

  • string_to_timestamp now outputs the datatype that is set as value timestamp_default_dtype

Snowflake

  • Fixed issue with multi-source hubs which had different business key names

Google BigQuery

  • Fixed issue with duplicate column names when using an input column as rsrc or ldts
  • Fixed uniqueness issue for Multi Source reference hubs

Synapse

  • Fixed issue with multi-source hubs and links
  • Fixed unnecessary quoting in PIT for the pit_type
  • Remove unnecessary subquery inside the ref_table

Exasol

  • Fixed Issues with Multi Active Hashing
  • Fixed several issues with column escaping
  • Added disable_hwm parameter to various RDV models, to align with other adapters