tzphh opened a new pull request, #10688: URL: https://github.com/apache/paimon/pull/10688
### Purpose Fixes #10658 Allow primary-key tables to maintain independent vector indexes on multiple columns through `pk-vector.index.columns`, supporting records with image, text, or other embeddings in the same table. Only columns listed in `pk-vector.index.columns` are indexed. Other VECTOR columns remain available for normal storage and reads without an ANN index. - Validate each configured column and preserve its own dimension, ANN implementation, distance metric, and index options. - Create and restore one vector index maintainer per configured column within each bucket. - Coordinate commit preparation, rollback, executor replacement, and resource cleanup across maintainers. - Reuse the existing field-specific query routing and index metadata format, and update the documentation. ### Tests - Core/Common: 85 tests passed. - Spark 3: 13 tests passed. - Flink 1.20.1: 11 tests passed. - Spotless, Checkstyle, Enforcer, Apache RAT, and `git diff --check` passed. Coverage includes field isolation, different dimensions and metrics, updates/deletes/nulls, compaction, recovery, failure rollback, and distributed searches across buckets. Tests use the deterministic test vector index backend. Native ANN integration tests and performance benchmarks were not run. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
