renatocron commented on issue #12986: URL: https://github.com/apache/druid/issues/12986#issuecomment-1294983860
Hello @abhishekagarwal87 I can't really reproduce it easily, it happens randomly depending on the filters I choose The original issue as open with a daily rollup, when I I open this issue I re-imported ~ 30 days of events from kafka, and during the ingestion while the segments were created, I had a few invalid results, then when I stopped the overload and run the query again, the results were OK, then I let auto-compaction runs with type=range and the issue appeared, then when I manually re-indexed with 'hash' the issue was gone This week, with the datasource I posted is not a rollup, and is using dynamic auto-compression, with PT72H delay, and my query is using `where __time > CURRENT_TIMESTAMP - interval '24' hour` so I think we can skip auto-compaction from the equation Note that this week issue only appeared when I added `TIMESTAMP_TO_MILLIS` to the `max(__time)` expression, when I run this only with `max(_time)` I did not see any invalid result, and now looking over, the original issue disappears when I added latest or array_agg and this week query is using LATEST as expression, so maybe it's not the same bug -- 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]
