Release Notes

August 2026

Resolves a BigQuery error when the segment, campaign, and automation sources supply mismatched datatypes by casting the segment_id join keys in mailchimp__segments to string.

Migrates the union_connections, apply_source_relation, and partition_by_source_relation macros to the dbt_fivetran_utils package.
Adds the fivetran_using_source_casing variable for case-sensitive destination support. When enabled, downstream transformations respect source casing to ensure consistent results. See the Additional Configurations section of the README for details.

Removed the dependency on the fivetran/mailchimp_source package.
- All functionality from the source package has been merged into this transformation package for improved maintainability and clarity.
- If you reference fivetran/mailchimp_source in your packages.yml, you must remove this dependency to avoid conflicts.
- Any source overrides referencing the fivetran/mailchimp_source package will also need to be removed or updated to reference this package.
- Update any mailchimp_source-scoped variables to be scoped to only under this package. See the README for how to configure the build schema of staging models.
As part of the consolidation, vars are no longer used to reference staging models, and only sources are represented by vars. Staging models are now referenced directly with ref() in downstream models.

Updated package to maintain compatibility with dbt-core versions both before and after v1.10.6, which introduced a breaking change to multi-argument test syntax (e.g., unique_combination_of_columns).
Temporarily removed unsupported tests to avoid errors and ensure smoother upgrades across different dbt-core versions. These tests will be reintroduced once a safe migration path is available.
- Removed all dbt_utils.unique_combination_of_columns tests.