SEPURI-SAI-KRISHNA commented on issue #12975: URL: https://github.com/apache/seatunnel/issues/12975#issuecomment-6410678740
Thanks @SEZ9. Both asks, in order. **1. The full comparison, and the test now drives it.** You offered two options and the first is the better one, so the locale switching test exercises all ten text bearing patterns instead of three, with `de-DE` added as a fourth default. Measured on `2024-06-15`, a Saturday in Q2, afternoon: | pattern | ENGLISH | tr-TR | ja-JP | de-DE | | --- | --- | --- | --- | --- | | `G` | AD | MS | 西暦 | n. Chr. | | `MMM` | Jun | Haz | 6月 | Juni | | `MMMM` | June | Haziran | 6月 | Juni | | `LLLL` | June | Haziran | 6月 | Juni | | `EEE` | Sat | Cmt | 土 | Sa. | | `EEEE` | Saturday | Cumartesi | 土曜日 | Samstag | | `ccc` | Sat | Cmt | 土 | Sa | | `a` | PM | ÖS | 午後 | PM | | `QQQ` | Q2 | Ç2 | Q2 | Q2 | | `QQQQ` | 2nd quarter | 2. çeyrek | 第2四半期 | 2. Quartal | All ten are now asserted under all four defaults, so the claim is checked by CI rather than by a table in a thread. **A correction to my earlier comment.** I said `Locale.ROOT` changes 2 of 7. Measured properly it is **4 of 10**: `MMMM`, `EEEE`, `QQQQ` and `LLLL` all abbreviate under ROOT. `Locale.ENGLISH` changes **0 of 10**. The conclusion is unchanged and in fact stronger, but the numbers I gave you were understated and I should have measured before quoting them. **2. Already in the PR.** `docs/en/transforms/sql-functions.md` and the ZH page both carry the behaviour note, under `FORMATDATETIME` and again under `FROM_UNIXTIME`. One thing you did not ask for, which came out of @DanielLeens's review and is worth knowing before you sign off: the pinned locale also pins `WeekFields`, so the localised numeric fields `Y`, `w`, `W`, `e` and `c` change too. A `de-DE` or `en-GB` worker formatting `2021-01-01` with `YYYY` got `2020` and now gets `2021`. `en-GB` matters there: an English language default is not interchangeable with `Locale.ENGLISH` for these fields. Both upgrade notes and both function pages now document it, and there is a regression covering it. -- 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]
