zhuxiangyi commented on PR #10394: URL: https://github.com/apache/paimon/pull/10394#issuecomment-6160253309
Thanks @JingsongLi, confirmed and fixed. The candidate rows were passed down as row ranges, which also decide whether `readTarget` reads the row sidecar, so filtered reads started to select it. They now restrict the selection the group already shares for the bitmap index (#10289), and the row ranges are unchanged, so this PR no longer affects sidecar selection. `testWideTableWithRowSidecar` covers your setup and fails for 7, 8, 15 and 16 columns before the fix. The decoding bug itself also exists on master: a `withRowRanges` read with the pushdown disabled returns the same shifted values. #10415 fixes it (verified on 1369458e78), so I left it there. -- 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]
