rubenada commented on PR #3665:
URL: https://github.com/apache/calcite/pull/3665#issuecomment-2528256328

   Sorry but the proposed approach (storing "originalRowCount" on the BNLJ) 
seems to go against some of the concepts discussed [on the 
Jira](https://issues.apache.org/jira/browse/CALCITE-6236?focusedCommentId=17813606&page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#comment-17813606):
   ```
   pre-storing selectivity or other kind of metadata information in the 
operator also feels like we are going to the wrong direction
   ```
   
   This is certainly a tricky one. IMO if we can't reach consensus about this 
topic, it'd not be a critical problem since each downstream project can always 
define their own `MyRelMdRowCount` handler and put in there whatever logic the 
require to refine BNLJ rowCount estimation.
   


-- 
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