taowenjun commented on issue #13087:
URL: https://github.com/apache/seatunnel/issues/13087#issuecomment-6571530884

   @SEZ9 I checked these points against the current implementation in #13085:
   
   - The English and Chinese Hive sink docs already describe the default-off 
option, precedence, whitespace/case normalization, missing/blank-property 
fallback, and unsupported-codec errors. They also explain that inherited ZSTD 
is rejected while its ORC mapping remains NONE; ORC/ZSTD support is handled 
separately in #13084, so #13085 alone does not enable inherited ZSTD.
   - Existing `HiveSinkConfigTest` coverage includes `" snappy "` 
normalization, explicit NONE overriding SNAPPY, an explicit codec overriding an 
invalid table property, missing/blank properties, non-ORC isolation, unknown 
codecs, and enum values without a supported ORC mapping. The 
configuration/options/factory tests passed (31 tests, no failures/errors/skips 
per JDK 11/17 Windows/Linux unit-test job) in [fork CI at implementation commit 
511c69c0b](https://github.com/taowenjun/seatunnel/actions/runs/41378715037). 
That run also executed `HiveIT#testExistingOrcTableCompression`, verifying 
inherited SNAPPY and explicit NONE through ORC footer and exact three-row data 
assertions. Both E2E build matrices used a JDK 11 engine container; this is not 
evidence for a JDK 17 engine runtime.
   
   On the scope question: **the inheritance option is ignored in both cases, 
rather than failing the job merely because it is enabled.** Non-ORC formats 
return before reading/validating `orc.compress`. When the table is absent at 
sink initialization, the automatic-creation path returns before inheritance; 
this also means `orc.compress` in `save_mode_create_template` is not inherited 
into the file writer by this option. Explicit `compress_codec`, existing 
defaults, and normal configuration/table-creation validation still apply.
   
   I have prepared an additional English/Chinese documentation clarification 
stating those two cases explicitly. This wording-only follow-up is currently 
local pending review/push; the behavior and tests described above are already 
in the published PR.


-- 
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