DanielLeens commented on issue #12935: URL: https://github.com/apache/seatunnel/issues/12935#issuecomment-6315603215
Thanks for the focused reproduction and for opening #12936. Classification: F / Connector-V2 File CSV reader semantics. The current `CsvReadStrategy` constructs the header-aware `CSVParser` before it consumes `skip_header_row_number`, so the preamble can indeed be parsed/buffered before the physical-line skip runs. That makes this distinct from the earlier header/double-skip cases: the requested behavior is to skip a non-CSV preamble before the real CSV header when file splitting is disabled. #12936 is the single implementation lane. Its bounded ordering change, reader regressions, LocalFile Zeta E2E coverage, and EN/ZH documentation are the right shape. Please keep follow-up implementation and test discussion on that PR, preserve the existing no-re-skip behavior for byte-range splits, and let the GitHub checks/review complete there. No parallel patch is needed; this issue stays open until #12936 is merged. -- 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]
