Dandandan commented on issue #956:
URL: 
https://github.com/apache/arrow-datafusion/issues/956#issuecomment-987690764


   Thanks for picking this up @ic4y 
   
   If you are getting a 10% improvement, this is already a fine  achievement 
(if we don't slow down the low cardinality too much?)
   
   The storage is one part of the story, as @alamb says it will require some 
changes in other places.
   
   Some ideas here:
   
   * Do not `slice` in the implementation, but rather append to aggregates 
directy (enabled by the change listed here)
   * More efficient (and vectorized?) implementation of `eq_array` based on 
typed arrays.
   * Avoid converting values into `ScalarValue` but use values and arrays 
directy.


-- 
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]


Reply via email to