cloud-fan commented on PR #43261: URL: https://github.com/apache/spark/pull/43261#issuecomment-1864013781
Just FYI, we hit a 10x perf regression with this change in our internal benchmark with a flattened directory layout. This PR says it can replace many recursive FileSystem calls, which is not an issue for flattened directories. Shall we disable it by default? Or we only use it when scanning a file source table and the table metadata says it has more than one partition column. -- 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]
