yihua opened a new issue, #905:
URL: https://github.com/apache/hudi-rs/issues/905

   ### Feature Description
   
   A record-level index lookup decodes every entry of every index block it 
opens into an owned record (a copy of the key and a copy of the value) and only 
then checks whether the key was requested. On a table with millions of keys the 
touched blocks hold millions of entries for a few thousand requested keys, and 
a profile of a 10,000-key upsert on a 10M-row table puts about 90% of its time 
in that decode.
   
   Proposed: walk the block's entries in place, compare the key bytes against 
the requested keys, and copy out only the entries that match.
   


-- 
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