WangGuangxin commented on pull request #35460: URL: https://github.com/apache/spark/pull/35460#issuecomment-1041421374
> If there is a fetch failure and the parent stage is `INDETERMINATE`, both the parent and child stage are recomputed. Custom RDD's can extend `getOutputDeterministicLevel` and return the right `DeterministicLevel`. See #22112 for more details Thanks for your reference. That's what I do in this MR for SparkSQL. When we shuffle by rand with sqls like `distributed by rand`, the RDD's DeterministicLevel generated by SparkSQL is `INDETERMINATE` after this MR -- 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]
