All releases
dbt Corev1.10
What's New in dbt Core 1.10
dbt Core 1.10 focuses on governance and observability: new contracts enforcement, richer metadata output, and a significantly improved dbt Mesh experience for large multi-project setups.
What's New
- 1Model contracts enforcement — define column-level types and constraints; dbt fails at compile time if they're violated
- 2dbt Mesh improvements — cross-project refs now support version pinning and deprecation notices
- 3New metadata output format — richer run_results.json and manifest.json for downstream tooling
- 4microbatch incremental strategy promoted to stable — large time-series tables now first-class citizens
- 5Unit tests support mocking of seed files and external sources
- 6Improved --select syntax with semantic layer metric traversal
- 7Python model improvements — better pandas/PySpark interop and type annotations
Breaking Changes
- dbt-core no longer bundles adapters — install dbt-postgres, dbt-snowflake, etc. separately
- run_results.json schema version bumped to v6 — update any tools parsing this file
Upgrade Notes
The adapter bundling change is the most impactful. Projects using dbt-core without specifying an adapter will fail on install. Add your adapter explicitly to requirements.txt.