voonhous opened a new pull request, #19916:
URL: https://github.com/apache/hudi/pull/19916

   ### Describe the issue this Pull Request addresses
   
   `TestLanceDataSource#testBlobOutOfLine` writes two rows pointing at `[0, 
1024)` and `[0, 512)` of one external file. A blob is a distinct entity and two 
blobs never share bytes (#18098), so `BatchedBlobReader` rejects the pair with 
`Overlapping blob ranges detected` whenever both rows land in the same Spark 
task. With the test harness at four shuffle partitions they never did. #19906 
moved the harness to two partitions and all six cases failed on Azure ([build 
17015, log 
98](https://dev.azure.com/apachehudi/a1a51da7-8592-47d4-88dc-fd67bed336bb/_apis/build/builds/17015/logs/98)).
 Context in #19911.
   
   ### Summary and Changelog
   
   - `blob_file_2.bin` grows from 1024 to 1536 bytes. Row 4 now references 
`[1024, 1536)`, disjoint from row 3's `[0, 1024)`. The expected offset for row 
4 follows.
   - No production code change.
   
   ### Impact
   
   Test only. The Lance out-of-line test no longer depends on which rows share 
a task.
   
   ### Risk Level
   
   none
   
   ### Documentation Update
   
   none
   
   ### Contributor's checklist
   
   - [x] Read through [contributor's 
guide](https://hudi.apache.org/contribute/how-to-contribute)
   - [x] Enough context is provided in the sections above
   - [x] Adequate tests were added if applicable
   


-- 
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