foxtail463 opened a new pull request, #69436: URL: https://github.com/apache/doris/pull/69436
### Problem Summary: Partition pruning already reflected partition predicates in the selected-partition row count, but filter estimation re-applied the same predicates as ordinary selectivities and undercounted cardinality. The recorded pruning proof was also matched back to scan outputs by table name, so rewrites that rebuilt slots could silently invalidate predicate removal. ### Solution: Record which partition predicates are already reflected in the scan row count, use them in filter estimation only to constrain column domains while preserving the row count, estimate only the remaining predicates normally, and estimate selected partitions without reported counts from the table rows not covered by known selected partitions. Carry the pruning proof on the scan so later rewrites either rebind it to the scan's slots or discard it, making the proof valid by construction. -- 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]
