#SQL

#PostgreSQL 19 Beta review uncovered a large number of unresolved, critical design and implementation flaws in the newly developed #SQL/PGQ property graph feature, ranging from broken dependency handling, insufficient locking mechanisms, incorrect permission checks to various crashes and unexpected runtime errors, that surfaced after a broad manual and AI-assisted code review.

Given the severity and volume of these unaddressed issues, combined with the upcoming release timeline for #PostgreSQL19, the #PostgreSQL Release Management Team reached a consensus with core contributors that the feature was nowhere near production-ready.

As a result, the community ultimately decided to fully revert the entire #PGQ implementation from both the #PostgreSQL 19 stable branch and the main development master branch, to eliminate time pressure and allow developers to rework the feature properly from scratch without being constrained by backwards compatibility requirements.

Learn more: https://www.postgresql.org/message-id/flat/dqa5mstx5mna3i7s23pdwl4m6bek7gqsgfccef44wjpswizufi%403aa6vzri3cat

The #PostgreSQL team has serious concerns about unresolved #SQL/PGQ design and dependency issues in #Postgresql 19, including orphaned metadata, inconsistent label/property behavior, and pg_dump problems. They recommend reverting #SQL/PGQ from #PostgreSQL 19 so the issues can be properly resolved for #PostgreSQL 20 without release and compatibility pressure.:

Improving #SQL Generation on Industrial Time-Series Data with Structured EXPLAIN Feedback using a 4 B SLM

Improving #SQL Generation on Industrial Time-Series Data with Structured EXPLAIN Feedback using a 4 B SLM