turboFei edited a comment on issue #25795: [SPARK-29037][Core] Spark gives duplicate result when an application was killed URL: https://github.com/apache/spark/pull/25795#issuecomment-532191422 I think it is feasible to detect concurrent writes. As known, when dynamicPartitionOverwrite is true, the job attempt path is `.spark-staging-${UUID}`. when dynamicPartitionOverwrite is false, it is `_temporary`. We just need check the existences of these paths and throw exception to fast fail when concurrent writes are detected. For that @advancedxy is working on [SPARK-28945](https://issues.apache.org/jira/browse/SPARK-28945) to enable concurrent dynamicPartitionOverwrite, should we thrown exception when detect the concurrent writes below? - dynamicPartitionOverwrite and dynamicPartitionOverwrite - dynamicPartitionOverwrite and staticPartitionOverwrite/non-partitioned table write @advancedxy @cloud-fan
---------------------------------------------------------------- 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. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
