2010YOUY01 opened a new issue, #17547:
URL: https://github.com/apache/datafusion/issues/17547

   ### Is your feature request related to a problem or challenge?
   
   This issue is a summary for https://github.com/apache/datafusion/issues/17488
   
   After [rewriting the NLJ 
operator](https://github.com/apache/datafusion/pull/16996/files), it has 
regression on the workload that the right side is very small (NLJ's left input 
has 1000 rows, right input has 1 rows)
   
   I think making it fast is good to have, but not necessary if it would 
introduce too many extra implementation complexity. 
   The reason is:
   - The optimizer will usually ensure the left side is smaller
   - Significant regression only happens when the right side is very small like 
only 1 row, otherwise even the optimizer make the suboptimal decision of join 
order, the performance should be similar
   
   See the original issue for more background
   
   ### Describe the solution you'd like
   
   _No response_
   
   ### Describe alternatives you've considered
   
   _No response_
   
   ### Additional context
   
   _No response_


-- 
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.

To unsubscribe, e-mail: [email protected]

For queries about this service, please contact Infrastructure at:
[email protected]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to