cloud-fan commented on pull request #34916: URL: https://github.com/apache/spark/pull/34916#issuecomment-995638178
> limit scene which data is less is not prone occur OOM We can't rely on assumptions. Technically the limit can be arbitrarily big. So I think writing to disk and reading back (shuffle) is better than collecting to memory and re-distribute. BTW AQE can optimize shuffles much better. -- 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]
