tustvold opened a new issue, #7518:
URL: https://github.com/apache/arrow-datafusion/issues/7518

   ### Is your feature request related to a problem or challenge?
   
   `BuiltInWindowExpr` currently makes use of the vectorised partition function 
when evaluating window frames with ranks, but falls back to a very inefficient 
mechanism based on performing scans using `Vec<ScalarValue>` as a comparison 
key for window functions that rely on the window frame, including user defined 
window functions.
   
   ### Describe the solution you'd like
   
   I may be missing something but it looks like window functions are only 
evaluated on data that is already sorted, as such it should be possible to 
identify the partition boundaries in a single pass, and then use this 
information to efficiently extract the window boundaries.
   
   Tagging @mustafasrepo and @alamb who may have more context on this design
   
   ### Describe alternatives you've considered
   
   _No response_
   
   ### Additional context
   
   #7517 tracks remove ScalarValue's implementation of Ord to cleanup the code 
and discourage this sort of approach in future


-- 
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]

Reply via email to