parthchandra commented on PR #5331: URL: https://github.com/apache/datafusion-comet/pull/5331#issuecomment-5689037370
Thanks for thanks for building the fork and measuring this. I verified this PR against against a personal build of Spark 3.4.3 + Iceberg build whose isOrderingEnabled is correct, which is why I didn't hit the nondeterminism you saw (your window/join getting wrong Spark answers is an Iceberg-side issue, not this PR). You're also right that the merge tests were on unpartitioned tables. I've moved all of them to the partitioned-by-constant pattern with assumeOrderingReported after the correctness check, so a build that reports ordering proves the merge ran. The above-the-cap sort fallback is now partitioned too. And the ascending/descending merge tests now verify order with a ROW_NUMBER() window (no ORDER BY to repair it), matching the existing window test. PTAL. -- 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]
