What's Changed

• fix: dry-run crash for invalid reference check in soda contract test by @m1n0 in #2604
• fix: stop fetching datasource config from Cloud in local verification flow by @m1n0 in #2606
• fix: handle metric query returning zero rows gracefully by @m1n0 in #2609
• fix: Update cleanup script by @Niels-b in #2610

Full Changelog: v4.1.0...v4.1.1

What's Changed

• fix: Potential issue with types not being converted to str in QueryResult by @Niels-b in #2590
• feat: Report missing count on duplicate and validity checks by @m1n0 in #2588
• fix: Snowflake connector for private key connections by @Niels-b in #2594
• fix: Add uv lock to release script by @Niels-b in #2601
• feat!: Separate SQL dialect from data source by @mivds in #2600
• feat: Expand structured data types and type casting to all applicable check types by @m1n0 in #2599
• feat: Add percentage-based sampling for postgres by @mivds in #2596
• feat: Add percentage-based sampling for databricks by @mivds in #2597
• feat: Add percentage-based sampling for snowflake by @mivds in #2598
• fix: Map COMPLETED_WITH_WARNINGS to WARNED in agent verification path by @m1n0 in #2605
• feat: Expand structured data types and type casting to reference check by @m1n0 in #2602
• chore: Test a single python version on PRs by @mivds in #2603
• fix: Apply column expression on freshness check correctly by @m1n0 in #2607
• fix: Update Fabric support by @Niels-b in #2595
• fix: Small fixes for Synapse data source by @Niels-b in #2608
• feat: Trino contracts support by @paulteehan in #2585

Full Changelog: v4.0.9...v4.1.0

What's Changed

• fix: migrates us fully to UV, support for PIP still maintained by @AidaPaul in #2577
• feat: Update Redshift for psycopg3 by @Niels-b in #2580
• feat: Support structured data types and type casting by @m1n0 in #2582
• fix: report correct datasetRowsTested for failed rows query check by @m1n0 in #2586
• fix: setup column expressions correctly for all aggregate metrics by @m1n0 in #2587
• fix: update failed rows check diganostics pyload by @m1n0 in #2592

Full Changelog: v4.0.7...v4.0.9

Please note: 4.0.8 failed during release pipeline and is not published in pypi.

This release failed during the pipelines. Please use 4.0.9.

What's Changed

• fix: migrates us fully to UV, support for PIP still maintained by @AidaPaul in #2577
• Update Redshift for psycopg3 by @Niels-b in #2580
• feat: Support structured data types and type casting by @m1n0 in #2582
• fix: report correct datasetRowsTested for failed rows query check by @m1n0 in #2586
• fix: setup column expressions correctly for all aggregate metrics by @m1n0 in #2587
• fix: update failed rows check diganostics pyload by @m1n0 in #2592

Full Changelog: v4.0.7...v4.0.8

What's Changed

• fix: Databricks - quote dataset in metadata query by @m1n0 in #2578
• Chore: Use tbump to generate github release by @mivds in #2575

Full Changelog: v4.0.6...v4.0.7

What's Changed

• Feat: Add uv support to soda core by @AidaPaul in #2561
• Feat!: Update to psycopg3 by @Niels-b in #2558
• Feat: add iterator support for query results by @mivds in #2562
• Fix: enhance CI workflow with staggered uploads to PyPI to reduce 503 errors by @nik843 in #2556
• Fix: Require pydantic>=2.11 by @paulteehan in #2559
• Fix: Improved Databricks data type mapping by @Niels-b in #2558
• Chore!: Replace strings with SamplerType enum by @mivds in #2565
• Chore: better yaml parsing for Soda Cloud client, support migration e… by @m1n0 in #2566

New Contributors

• @nik843 made their first contribution in #2556
• @AidaPaul made their first contribution in #2561

Full Changelog: v4.0.5...v4.0.6

This is the first public release of Soda Core version 4. This release introduces Data Contracts as the default way to define data quality rules for tables. The new approach offers a cleaner, more structured, and more maintainable way to define and manage data quality rules, based on community feedback and real-world usage.

Breaking change: Soda Core is moving from the checks language to a Data Contracts–based syntax.

Highlights

• Introduced support for parsing, publishing and (both locally and remotely) verifying data contracts.
• Introduced extended check diagnostics providing deeper visibility into the count and percentage of tested, passing, and failing rows.
• Introduced seamless flexibility to run Contracts via Core or the Soda Agent, with definitions stored either as external files or in Soda Cloud.
• Introduced "Missing", "Invalid", "Duplicate", "Aggregate", "Failed Rows", "Metric" checks in data contract.
• Introduced data contracts support for multiple data sources: Postgres, Snowflake, BigQuery, Databricks, Redshift, SQL Server, Fabric, Synapse, Athena, DuckDB (in memory).
• Introduced new CLI with a noun-verb structure and better integration with the Soda Cloud APIs (e.g. contract fetching based on a dataset identifier).
• Introduced support for variables in contracts, allowing you to parameterize contracts.
• Introduced support for extending functionality using plugins.
• Introduced extensible check types.
• Introduced concept and hooks for contract verification result handlers, allowing post-processing of the contract verification results.

Extensions

• Implemented the contract generation plugin.
• Implemented the diagnostics warehouse plugin.
• Implemented reconciliation checks.
• Implemented group-by checks.
• Implemented support for Oracle as a data source.
• Implemented support for Dremio as a data source.
• Implemented support for contract requests.

What's Changed

• Update readme callout with soda cloud call to action by @santiviquez in #2366
• feat(snowflake): add oauth support by @adkinsty in #2415

Full Changelog: v3.5.5...v3.5.6

What's Changed

• Update README.md with launch banner by @santiviquez in #2292
• Fix authentication inside Fabric Notebooks by @sdebruyn in #2299
• Add dotenv to deps, fixes #2285 by @m1n0 in #2312

New Contributors

• @santiviquez made their first contribution in #2292

Full Changelog: v3.5.4...v3.5.5

Full Changelog: v3.5.3...v3.5.4