SEZ9 commented on issue #12975: URL: https://github.com/apache/seatunnel/issues/12975#issuecomment-6393361939
Thanks for the detailed report and for measuring against the real `DateTimeFunction` methods rather than a reimplementation. The contrast with `DAYNAME`/`MONTHNAME`, which already pin `Locale.ENGLISH` via `DateFormatSymbols`, is a strong sign the locale-dependent output of `FORMATDATETIME` and `FROM_UNIXTIME` is unintended. A fix would likely be to pin an explicit locale when building the formatter in both `formatdatetime` and `fromUnixTime`, plus a unit test that runs under a non-English default locale to guard against regressions. Since you've indicated you're willing to submit a PR, please go ahead. One thing to keep in mind: pinning to English is a behavior change for users on non-English workers who currently get localized month/day names, so it's worth calling out in the PR description. An optional locale argument could be considered in a follow-up, but deterministic output across workers should be the default. <!-- streview-comment:2024 --> -- 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]
