yihua opened a new pull request, #19384:
URL: https://github.com/apache/hudi/pull/19384

   ### Describe the issue this Pull Request addresses
   
   Promoting a bare `long` column to a timestamp logical type during 
writer-schema deduction is not uniformly gated: `long -> local-timestamp` is 
authorized only by an explicit per-field override, but `long -> UTC-timestamp` 
(`timestamp-micros` / `timestamp-millis`) is silently allowed on the default 
write path. A bare long carries no precision signal (millis vs micros), so 
silently attaching a UTC timestamp logical type can mislabel stored values.
   
   ### Summary and Changelog
   
   Gate `long -> timestamp` promotion uniformly behind the per-field override 
for all four target types (UTC and local), in both reconcile paths: rejected 
with an actionable error without an override, applied with one.
   
   Placeholder draft. The implementation depends on the per-field 
timestamp-logical-type override framework and the InternalSchema timestamp type 
extensions, which are not yet upstream. Those dependency PRs will be filed 
first; the code and tests will be added here once they land.
   
   ### Impact
   
   A write that attaches a timestamp logical type to an existing bare-long 
column requires a per-field override; without one it fails loud instead of 
silently applying.
   
   ### Risk Level
   
   low
   
   ### Documentation Update
   
   none
   
   ### Contributor's checklist
   
   - [ ] Read through [contributor's 
guide](https://hudi.apache.org/contribute/how-to-contribute)
   - [x] 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