HyukjinKwon commented on pull request #30167: URL: https://github.com/apache/spark/pull/30167#issuecomment-717897756
@RussellSpitzer, I completely agree with this fix itself. But about porting it back: The flip side works too. For example, for any reason, in a cluster the catalog was set and users used to use it with no problem. User don't want to change their existing workload and code but switch the catalog they are using. So they just reset the configuration from their cluster. After the fix, this case wouldn't work anymore. Of course, I like this failfast way so it can prevent such misuse but from user's perspective I'd imagine they would think a behaviour change which should ideally not happen in maintenance release. To be more clear, - if users set their catalog properly, things work correctly with/without this fix. - this fix throws an exception for the case (we defined as misuse now) which might have worked before I think it's arguable to call it a bug fix to port back. ---------------------------------------------------------------- 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]
