Dandandan commented on issue #3214:
URL: 
https://github.com/apache/arrow-datafusion/issues/3214#issuecomment-1230581186

   > Might `count(*)` be as simple as a stats lookup in Parquet or DeltaLake? 
Reading a billion values just to count them seems sub-optimal, but that can 
definitely be addressed with a TODO and a future PR.
   
   You're right, for a schema provider that has statistics available, we can 
skip scanning.
   AFAIK DataFusion already has support for using the statistics-provided 
count/min/max from the provider (e.g. delta lake).
   
   You're right that we could also use the parquet statistics for files instead 
of skipping reading the columns. I think we don't support this yet. At least 
for min/max statisticd his avoids  having to scan the entire column and compute 
the min/max.


-- 
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