HyukjinKwon edited a comment on pull request #28745: URL: https://github.com/apache/spark/pull/28745#issuecomment-641674305
> will this cause perf regression? e.g. if grouping expr is expensive, with the Project we only need to evaluate it once. I would say this is kind of a design choice. In the other way, we should add the projection to all grouping expressions, and will need to keep more data intermediately on the other hand. This PR matches the implementation with existing grouping expressions - it shouldn't be matched with object expressions because grouped and cogrouped UDFs actually should pass a key separately to UDF to use which object expressions don't. ---------------------------------------------------------------- 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] --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
