rangareddy commented on issue #17041:
URL: https://github.com/apache/hudi/issues/17041#issuecomment-5391450243

   This issue was reviewed as part of the JIRA-migrated backlog triage 
(HUDI-9496).
   
   **Findings: needs a reproduction - the only evidence attached is a 
screenshot.**
   
   The report is clear in substance: decimal partition values are written 
incorrectly into the filesystem **path** while being correct in the data file, 
which matters for Trino specifically because the connector derives partition 
values from the path rather than the data.
   
   The obstacle is that the evidence is an inline image (`!image.png!`), so the 
exact wrong-versus-right values are not recoverable from the ticket text. Could 
the actual values be pasted in as text? That alone would make this directly 
workable.
   
   To reproduce: write a table with a decimal (and separately a timestamp) 
partition field, then compare the directory names on storage against the values 
in the data file.
   
   **Four tickets in the backlog are the same underlying theme** - "the 
partition value written to the path is not faithful to the value" - and are 
worth looking at as one problem:
   
   - this one - decimal/timestamp partition values wrong in the path
   - #17003 (HUDI-9396) - Flink's path-based partition-spec parsing mis-splits 
a value containing `/`
   - #17224 (HUDI-4430) - incorrect type casting reading a table written with 
`CustomKeyGenerator` and a unixtimestamp partition field
   - #17247 (HUDI-7306) - `CustomKeyGenerator` truncating the microseconds part 
of a timestamp
   
   Keeping this open.
   


-- 
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