JingsongLi commented on PR #9896:
URL: https://github.com/apache/paimon/pull/9896#issuecomment-5708353706

   Closes a real gap. Level-0 APPEND files are not index sources, so filtering 
by `shouldRead` before the bucket search silently dropped freshly written rows 
from FULL/DETAIL vector searches — exactly the rows a user expects to find 
right after writing. Keeping the source policy in the state builder for ANN 
segments while passing all data files to the bucket search is the right split, 
and FAST still omits uncovered files. The two new tests cover both halves 
(found in FULL, still omitted in FAST).
   


-- 
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