PingLiuPing opened a new issue, #11023:
URL: https://github.com/apache/incubator-gluten/issues/11023

   ### Description
   
   Currently, Iceberg partition transform is performed in the execution layer 
(within TableWrite operator). For better architecture and optimization 
opportunities, this computation should be moved to the optimizer/planner layer 
where it logically belongs. And this can be achieved by adding a projection 
node above TableWriter node and letting this node to perform the partition 
transform and send the pre-computed partition keys to TableWriter.
   
   Doing so would allow for this computation to be optimized further by, for 
example, pushing it below the exchange and letting it run with higher 
concurrency than write itself. It would also provide more transparency for 
debugging as we would get runtime stats just for this piece of computation. It 
would allow users to see more clearly in the query plan what is happening 
allowing for more efficient debugging.
   
   ### Gluten version
   
   None


-- 
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