jayzhan211 commented on PR #26228: URL: https://github.com/apache/datafusion/pull/26228#issuecomment-6244057352
Could the NOT IN join use the null-aware RightAnti from #23957 instead? Then the broadcast subquery is the build side and Spark's streamed side is the probe side. The probe order is already declared through maintains_input_order, so Comet wouldn't need any build-side promise. It would also stop building the hash table from the whole streamed partition. The one limit is that it doesn't accept a join filter. -- 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]
