SteNicholas commented on pull request #2111: URL: https://github.com/apache/hudi/pull/2111#issuecomment-713335563
> > If user doesn't essentially care about duplicates, I agree that we need to have same behavior w/o small file handling. > > I don't follow this actually. I understand that this fixes the issue that an insert be treated as an update, if the user generates the same key. (but I think of it more as a misuse of key i.e its the user's responsibility to generate unique keys and this wont happen). Nonetheless, this might be a valid mode to have for users who can't bother to create unique keys. but like said above, it would be nice to retain small file handling, while just allowing duplicates through a separate IOHandle/flag? @vinothchandar I agree with the point that allowing duplicates through a separate IOHandle/flag. I would like to invoke a different IOHandle to HoodieMergeHandle to allow duplicates. ---------------------------------------------------------------- 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]
