fhan688 commented on issue #19286:
URL: https://github.com/apache/hudi/issues/19286#issuecomment-4967967191

   > For 0.x table timeline, the completion time is of an instant is its 
request time or instant time, the semantics are backward compatible.
   
   Agreed, and thanks — this is the right framing. You're correct that for a 
LAYOUT_VERSION_1 timeline an instant's completion time is its instant (request) 
time, and the semantics are backward compatible. In fact the V1 view already 
reflects this: load() maps requestedTime → completionTime, and 
getCompletionTime returns the begin time for archived/short instants.
   
   So instead of the current parallel getInstantTimes implementation in 
CompletionTimeQueryViewV1, I'll refactor PR1 to reuse the shared range logic 
driven by "completion time = instant time" for V1, letting V1 fall out as a 
special case rather than a separate code path. This removes the duplicated 
range/OPEN-CLOSED/resume handling and shrinks the regression surface.
   
   One boundary I'll keep and call out explicitly: V1 intentionally does not 
load the archived timeline, so for archived instants the completion time falls 
back to the instant time (which is exactly the equivalence you describe). I'll 
update the umbrella/PR wording accordingly (dropping the "have no completion 
time" phrasing in favor of "completion time = instant time"). Will push the 
refactor shortly.


-- 
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