tagadearpit opened a new pull request, #11950:
URL: https://github.com/apache/seatunnel/pull/11950
Fixes #11949
### Purpose of this pull request
This change makes the SQL transform `lower()` and `upper()` functions
deterministic across worker machines by using `Locale.ROOT` instead of the JVM
default locale. Under a Turkish default locale, the previous implementation
converted `I` to `ı` and `i` to `İ`, which could make identical jobs produce
different results on different workers.
The implementation is limited to `StringFunction.lower()` and
`StringFunction.upper()` and preserves existing null behavior.
### Does this PR introduce any user-facing change?
Yes. SQL case conversion is now locale-independent. For example, with the
JVM default locale set to `tr-TR`:
```text
Before: lower("I") = "ı", upper("i") = "İ"
After: lower("I") = "i", upper("i") = "I"
```
This is a correctness fix for deterministic SQL transformations and does not
require configuration changes or a migration.
### How was this patch tested?
A regression test was added to `StringFunctionTest` that temporarily sets
the default locale to Turkish, verifies the expected `lower()` and `upper()`
results, and restores the original locale in a `finally` block.
Verification completed locally:
- `StringFunctionTest`: 34 tests passed, 0 failures, 0 errors.
- SeaTunnel transforms module verification lifecycle: passed.
- SeaTunnel Spotless formatting/check: passed.
- Direct pre-fix harness reproduced `lower=ı upper=İ`.
- Direct post-fix harness passed with `lower=i upper=I`.
- `git diff --check`: passed.
The repository-wide `./mvnw -q -DskipTests verify` and `./mvnw test`
commands were also attempted. They are blocked by existing
repository/environment issues in the config-shade module and frontend build;
the changed transforms module and its complete `StringFunctionTest` pass
independently.
--
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]
