siddharthteotia commented on PR #15698:
URL: https://github.com/apache/pinot/pull/15698#issuecomment-2848466737

   Apologies if these questions have been discussed earlier. In that case, you 
can just point to the discussion or doc. 
   
   I have a couple of questions out of curiosity --
   
   - Are we primarily trying to optimize the JOIN ordering or broad set of 
queries ? 
   
   - Have we looked at LOptJoinOptimizeRule in Calcite ?
   
   - I don't see a cost factory. We also don't have a stats store. So is this 
rule/HEP based ?
   
   > FWIW I don't expect this new optimizer to add any new overheads of its own.
   
   I think it really depends on the query complexity. In my experience, 
sophisticated query planning can easily go upto 100s of ms or even upto 1sec 
but then then it could be worth it if still optimizes the execution. 
   
   I am very interested to know what are the kinds of queries for which you are 
seeing performance boost with the physical optimizer. 
   
   > that is temporary: usePhysicalOptimizer=true/false. By default this is 
assumed false. When we use the physical optimizer, we need to skip certain 
sections of the HepProgram and I have made those changes
   
   Why is this needed? I am hoping that even in the first working version of 
the optimizer, there is no query option otherwise we are just going to create a 
lot of tech debt / conditional paths where people will inevitably bypass it and 
create their own niche optimizations in the OSS code. 
   
   Overall I think this is a step in the right direction. It will make Pinot 
better and help make it feature complete. However, optimizer is a long journey 
(longer than execution) with lots of consistent non-trivial 
optimizations/revisions. I suggest we invest systematically with a long term 
direction which I am not sure what that is. If there is a doc, I would love to 
see it.


-- 
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: commits-unsubscr...@pinot.apache.org

For queries about this service, please contact Infrastructure at:
us...@infra.apache.org


---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@pinot.apache.org
For additional commands, e-mail: commits-h...@pinot.apache.org

Reply via email to