github-actions[bot] commented on code in PR #63167:
URL: https://github.com/apache/doris/pull/63167#discussion_r3232779498


##########
be/src/storage/delete/delete_bitmap_calculator.cpp:
##########
@@ -199,6 +199,18 @@ Status 
MergeIndexDeleteBitmapCalculator::calculate_one(RowLocation& loc) {
             break;
         }
         _last_key = cur_key.to_string();
+        if (_rowid_length > 0) {
+            // For cluster-key MOW tables, the primary key index contains 
rowid and

Review Comment:
   This fixes the seek skip, but it now preserves the first rowid encountered 
in a segment and deletes later rows with the same unique key. For cluster-key 
MOW the rowid suffix is encoded ascending (`SegmentWriter::_encode_rowid`), so 
duplicates within one segment are visited from smaller to larger rowid. When 
`_last_key` matches on the second duplicate, `calculate_one()` returns that 
later row as deleted, leaving the older row visible. The intended MOW behavior 
is to keep the newest row (larger rowid within the same segment, absent a 
higher sequence value), so this can still return stale data for exactly the 
same same-segment duplicate case this PR is trying to fix.



##########
regression-test/data/audit/test_audit_log_behavior.out:
##########
@@ -1,6 +1,6 @@
 -- This file is automatically generated. You should know what you did if you 
want to edit this
 -- !audit_log_schema --
-query_id       varchar(48)     Yes     true    \N      
+query_id       varchar(128)    Yes     true    \N      
 time   datetime(3)     Yes     true    \N      

Review Comment:
   This line introduces trailing whitespace, and `git diff --check` reports it 
as an error. Please regenerate or trim the expected output so whitespace checks 
pass.



-- 
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]


---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to