lokeshj1703 opened a new pull request, #19751:
URL: https://github.com/apache/hudi/pull/19751

   ### Describe the issue this Pull Request addresses
   
   closes #19750
   
   ### Summary and Changelog
   
   Partitioned record-level index (RLI) breaks for a single-level hive-style 
partition whose column NAME contains a dot (for example a column 
`fare.currency`, segment `fare.currency=USD`): the literal dot in the column 
name makes the metadata-table log file name unparseable by the log-file pattern 
(`InvalidHoodiePathException`).
   
   `PartitionPathEncodeUtils.escapeFileName` now, for a single-level hive-style 
segment, escapes dots in the column name (before the first `=`) while leaving 
dots in the partition value (after `=`) as-is. `TestPartitionPathEncodeUtils` 
covers the escaping.
   
   Scope and limitations (deliberately narrow): this handles only the 
single-level dotted-column-name case. Partition VALUES that contain dots (e.g. 
`date=2024.01.01`), nested hive-style partitions (e.g. 
`region=US/fare.currency=USD`, only the first `=` is handled), and 
non-hive-style dotted paths remain unhandled and still fail. #18378 takes the 
broader escape-all-dots approach that also covers those cases; that is the more 
complete fix. Opening this to compare the two.
   
   ### Impact
   
   `escapeFileName` escapes dots in single-level hive-style partition column 
names. Partition values are unchanged. Does not fully fix partitioned RLI for 
value-dots or nested/non-hive dotted partitions.
   
   ### Risk Level
   
   low. Confined to file-name escaping and covered by unit tests.
   
   ### Documentation Update
   
   none
   
   ### Contributor's checklist
   
   - [ ] Read through [contributor's 
guide](https://hudi.apache.org/contribute/how-to-contribute)
   - [ ] Enough context is provided in the sections above
   - [ ] 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