DanielLeens commented on issue #10129:
URL: https://github.com/apache/seatunnel/issues/10129#issuecomment-7008673530

   Classification: F / Connector-V2 File ORC stripe splitting.
   
   Thanks for the proposed follow-up. I rechecked current `dev` at 
`1d0e6509a0aba3e34219a1a7ae370913f0150b2a`: `FileSplitStrategyFactory` still 
routes Parquet to its dedicated strategy and otherwise uses byte-size splitting 
only for formats that advertise split support, while ORC still reads with a 
single `Reader.Options().schema(schema)` path. I found no current open PR for 
this ORC splitting scope, so the scope remains available.
   
   There are no new requirements beyond the earlier guidance. Please build the 
new strategy through the current shared `FileSplitStrategyFactory`, using the 
connector `HadoopConf` and `HadoopFileSystemProxy`, so LocalFile, HDFS, and S3 
keep one contract. Split only on complete ORC stripes; preserve the projected 
schema when adding the ORC range; and do not change `FileSourceSplit` 
serialization or legacy start/length restore semantics.
   
   The PR should include a stripe-boundary unit test, split-versus-unsplit 
projection equivalence, an existing-style LocalFile E2E with parallelism above 
one and exact row counts, plus matching English and Chinese documentation. 
Please credit the original #10325 contribution in the new PR and link it here 
when ready.


-- 
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