neilconway commented on PR #22343:
URL: https://github.com/apache/datafusion/pull/22343#issuecomment-4483681627

   > But also the only real win seems `Q73` in tpcds? What is your intuition 
for where the win is coming from? I'm wondering if it's just happening to hit a 
positive case that would be handled by #22144 already or if it's completely 
unrelated (e.g. in a complex join key).
   
   I see consistent improvements on ClickBench Q21 (~10-13%) and Q22 (~5%), 
which are both cases where we now reorder `LIKE` predicates after simple 
comparisons.
   
   Interestingly, this PR does *not* fire for Q73 in TPC-DS, so I'm not sure 
what is going on there 😊 I couldn't repro an improvement locally, so I guess it 
is just benchmark noise.
   
   To see improvements for a broader class of queries, we'd need to extend the 
heuristics to consider more criteria.


-- 
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