wombatu-kun commented on PR #8735: URL: https://github.com/apache/paimon/pull/8735#issuecomment-5018881150
Fixed in 36a07fa55. `GlobalIndexScanner` now attaches `rowCount` (the `[0, rowCount)` complement path) only when a row range is backed by a single index file; multi-file ranges fall back to the full-file scan, so their sparse, column-sorted row ids stay correct. Regression test added in `BtreeGlobalIndexTableTest#testMultiFileRangeNegationPredicatesUseAbsoluteRowIds`. A possible follow-up (keeping multi-file ranges fast via a range-level complement) is noted in the PR description. @JingsongLi @steFaiz -- 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]
