SEZ9 commented on issue #12571:
URL: https://github.com/apache/seatunnel/issues/12571#issuecomment-5987711312

   Thanks @SEPURI-SAI-KRISHNA, that closes the loop on the coverage question.
   
   What you describe at `c5b287cc5` matches what was asked for on this issue:
   
   - 
`SystemFunctionTest.testCastAsTinyintAndSmallintAlreadyRejectOutOfRangeNumericInput`
 pinning the existing `TINYINT` / `BYTE` / `SMALLINT` parsing paths 
(out-of-range rejected, in-range still converts, fractional source such as 
`5.7d` still rejected) is exactly the justification needed for leaving them off 
the shared numeric helper. Good to have that written down as a test rather than 
as a comment.
   - `testCastAsIntRejectsOutOfRangeNumericInput` covering the `INT` path 
directly, plus the correction so the range check compares the truncated value 
and `CAST(2147483647.5 AS INT)` is accepted, is the right fix for the 
first-version behaviour.
   - `ZetaSQLEngineTest.testCoalesceAndIfnullAlsoRejectAnOutOfRangeNarrowing` 
asserting the error for `3000000000L` instead of the previously wrapped 
`-1294967296`, with the in-range re-run on a fresh engine, is what I wanted for 
the `COALESCE` / `IFNULL` boundary. The comment explaining that the result type 
is inferred from the first non-null argument is useful context for whoever 
touches that later.
   - The incompatible-changes entry on both the en and zh pages, including the 
`TRY_CAST` migration guidance, covers the documentation ask.
   
   Nothing further from me on this thread's scope. Remaining asks, all on the 
PR side:
   
   1. Keep the three tests and the docs entry together in the same PR as they 
are now; please don't split them out if there is further rebasing.
   2. Once the Build on the PR has finished, drop a note there so review can 
proceed on a green run. I'll keep this issue open until the PR is in `dev`.
   
   <!-- streview-comment:1529 -->


-- 
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