IvanVergiliev commented on issue #24068: [SPARK-27105][SQL] Optimize away 
exponential complexity in ORC predicate conversion
URL: https://github.com/apache/spark/pull/24068#issuecomment-499470485
 
 
   @cloud-fan I applied the changes to the `hive/` subtree as well. It seems 
like it wasn't updated to reflect some of the recent changes to the other files 
(for example, it doesn't seem to have the quoted attribute name support) - I 
only applied the minimum amount of changed required to get the same algorithmic 
behaviour. I haven't applied the quoted attribute name support and other 
differences.
   
   @gatorsmile I see what you're saying. I got rid of the benchmark that only 
does the filter conversion step since that indeed seemed a bit out of place 
with the rest. The remaining new benchmark is `Pushdown benchmark with 
unbalanced Column` which tests the full filter-with-pushdown sequence that the 
rest of the benchmarks cover. This remaining benchmark is almost the same as 
`Pushdown benchmark with many filters` which was added in 
https://github.com/apache/spark/pull/22313 with the introduction of the 
`buildTree` function. I think it makes sense to have these two benchmarks next 
to each other. Let me know if you'd like me to move both of them to a new 
benchmark file. (I have a small concern that if we move them to a separate 
file, it'll be easier to forget to run them when making changes - because, as 
you say, the process is not fully automated at the moment. Thus, I slightly 
prefer keeping them where they are, but I'm open to moving them too.)

----------------------------------------------------------------
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]


With regards,
Apache Git Services

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

Reply via email to