[GitHub] [spark] planga82 commented on pull request #32057: [SPARK-34961][SQL] Migrate First function from DeclarativeAggregate to TypedImperativeAggregate to improve performance

[GitBox <gi...@apache.org>]

planga82 commented on pull request #32057:
URL: https://github.com/apache/spark/pull/32057#issuecomment-814406977


   Hi @tanelk ,
   First, thanks for sharing the benchmark, I was concerned about something like that could happen,it does not seem very reasonable to lose performance in this case. 
   
   Without having a very deep understanding, `ObjectHashAggregateExec` uses a different way to manage the aggregation buffer, it supports buffer states with variable sizes, so it needs it’s own iterator (`ObjectAggregationIterator`) and allows to define a buffer in the `typedImperativeAggregate` function. 
   
   Thank you for your input, i will try to find a solution that don’t lose the performance we have achieve with DeclarativeAggregate functions


-- 
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:
users@infra.apache.org



---------------------------------------------------------------------
To unsubscribe, e-mail: reviews-unsubscribe@spark.apache.org
For additional commands, e-mail: reviews-help@spark.apache.org