Caideyipi opened a new pull request, #19032:
URL: https://github.com/apache/iotdb/pull/19032

   ## Description
   
   LOAD can split batched aligned value chunks independently from their time 
chunks. The splitter previously dropped null row positions and empty pages, 
wrote decoded values with a dummy timestamp, and produced incorrect value 
statistics or unreadable page layouts. Preserve every row and page slot while 
rebuilding the standalone value chunk, and use the actual timestamp for value 
encoding and statistics.
   
   Add regression coverage for sparse null rows, copied empty pages, and 
decoded empty pages across time partitions. The cases write the batched 
compaction layout, split it, write each partition back, and verify metadata and 
query results.
   
   This change is independent of #19031, which covers LOAD snapshot and 
progress-log races.
   
   ## Validation
   
   - `mvn spotless:apply -P with-integration-tests`
   - Clean remote JDK 17 dependency reactor: the three new regressions passed 
(3 tests, 0 failures), and the full `TsFileSplitterTest` passed (15 tests, 0 
failures).
   - The previous implementation reproduced one missing-row failure and two 
`BufferUnderflowException` errors in the same three regressions.
   


-- 
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