szehon-ho commented on PR #58145: URL: https://github.com/apache/spark/pull/58145#issuecomment-5923062857
Addressed the remaining validation, coverage, and documentation findings in [fc995075e62](https://github.com/apache/spark/commit/fc995075e62b76fa7073dd753673344d4d95b193). - Reject source-reported struct/array field ordinals before rebinding, and reject unsupported `Unevaluable` / `RuntimeReplaceable` expressions while allowing column references. - Add regressions for both opt-in settings, fully pushed/inferred deduplication, and three-input scan rebuilding under both symmetric modes, including disabled controls. - Clarify SQL `true` and callback eligibility, remove the unused `ValuesReaderFactory` companion, use structured warning fields, and correct the canonicalization comment. All 44 focused test cases passed, along with Catalyst main and SQL main/test Scalastyle. Local validation used temporary LZ4 1.11.2 because the Maven proxy still lacks 1.12.0; 1.12.0 is restored and there is no dependency downgrade in the PR. Scan-metadata constraint propagation remains deferred. With `useInferredFilterEstimation() == false` and `reflectsFullyPushedDownFilters() == false`, inferred predicates remain ordinary `Filter` nodes and participate in the existing constraint machinery when enabled. -- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
