HyukjinKwon commented on pull request #28661: URL: https://github.com/apache/spark/pull/28661#issuecomment-636311879
That sounds more conservative way, yes. If it hides any useful information, I think that's the best way to do so. But for PythomUDF specifically here, I really think the stacktrace is useless because there is only one place in JVM side the exception is thrown. Also, this case is when it captures the failure from Python workers only. Other whitelisted exceptions, they already hide the JVM stacktrace in Python 2 and I dont think we have heard many complaints about it. If we'll do more general approach later, I kind of agree with your idea. But I would like to get this in separately for now if that sounds fine to you. ---------------------------------------------------------------- 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]
