Aggarwal-Raghav commented on PR #6170: URL: https://github.com/apache/hive/pull/6170#issuecomment-3501850629
> This fix makes sense to me overall — split waves is indeed a float, so truncating it to an int is clearly a bug. > > It’s worth noting that waves is buried deep inside the split generation logic, making it an expert-level setting. So it’s not surprising that this issue hasn’t been fixed until now (based on customer escalation history). > > mm_all.q.out change also makes sense, it's the opposite of that was done in [HIVE-19703](https://issues.apache.org/jira/browse/HIVE-19703) > > I wish we have a much cleaner unit test to reflect how these settings affect the number of splits, but it's not necessarily the scope of this PR Thanks for the reply @abstractdog . There is 1 UT failure `mm_dp.q`. I'm debugging that, seems relevant. I'll see if a junit test is possible for the test the number of containers launched logic. -- 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]
