SEZ9 commented on issue #12658: URL: https://github.com/apache/seatunnel/issues/12658#issuecomment-6029514588
Thanks for the detailed report and reproduction. Since #11264 already fixed the analogous stale-formatter cache in the JSON format, aligning TextDeserializationSchema and CsvDeserializationSchema with that approach (fall back to per-value pattern detection when the cached formatter fails, or cache per precision) sounds like the right scope. Please go ahead with the PR. In the PR, please include unit tests for both TextDeserializationSchema and CsvDeserializationSchema covering mixed precision in both orders (0 -> 3 -> 6 digits and reverse) and asserting nanoseconds, plus a negative case confirming unsupported text still throws. Please also check whether the same cached-formatter pattern exists for DATE/TIME fields or in other text-based formats (e.g., file source readers that share DateUtils/TimeUtils helpers) so the fix is consistent across formats. <!-- streview-comment:1569 --> -- 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]
