Dandandan edited a comment on pull request #9007: URL: https://github.com/apache/arrow/pull/9007#issuecomment-751466654
I think that definitely makes sense @alamb to test it like that, I think the second case (bogus stats) makes sense to start with, but to have a better look based on the TPCH queries makes sense too at some moment (I think we would need to support more queries first though that benefit from reordering). Currently we also only swap order `locally` between two nodes, at some point we would like to do have a more global optimization, making it even more important to have proper tests. ---------------------------------------------------------------- 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]
