DataJoint 2.2 Released

[dimitri-yatsenko]

We are happy to announce the release of DataJoint 2.2 for Python.

For a comprehensive overview, see What's New in DataJoint 2.2.

Highlights

Graph-driven cascade delete and restrict

The new Diagram.cascade() classmethod replaces the old error-driven cascade approach. Delete and drop operations now use the pipeline DAG to determine affected tables before executing. New methods: Diagram.cascade(), Diagram.restrict(), Diagram.prune(), Diagram.counts(), with full dry-run support via safemode=True. This also fixes the longstanding cascade delete failures on MySQL 8 (#1110).

Thread-safe mode with dj.Instance

New dj.Instance() class provides independent database connections with connection-scoped configuration, enabling safe concurrent access from multiple threads — essential for web servers, parallel workers, and multi-pipeline applications.

Directory references in filepath storage

<filepath@store> now supports directory references. is_dir is detected dynamically and existence checks handle directories correctly.

Additional fixes

• populate() with reserve_jobs=True now correctly applies restrictions (populate() with reserve_jobs=True ignores *restrictions #1413)
• Populate antijoin uses .proj() for correct pending key computation (Fix populate antijoin to use .proj() for correct pending key computation #1405)
• Attribute names starting with index no longer cause parse errors (Bug: Cannot create attributes with names starting with 'index' due to regex mismatch in declaration parser #1411)
• Backend-agnostic quoting and adapter abstractions for cleaner MySQL + PostgreSQL support (fix+refactor: backend-agnostic quoting, table name construction, and adapter abstractions #1419)
• skip_duplicates=True behavior on PostgreSQL documented and tested (Advanced handling of duplicates in insert1 #1049)

Install or upgrade

pip install datajoint==2.2.0

Full release notes: https://github.com/datajoint/datajoint-python/releases/tag/v2.2.0