jbonofre opened a new pull request, #1456: URL: https://github.com/apache/arrow-java/pull/1456
## What's Changed `dataset/src/test/resources/substrait/local_files_users.json` declared `ID` as `NULLABILITY_REQUIRED`. The Parquet file that `testRunQueryLocalFiles` reads has an optional `id` column, and the test expects a nullable `ID`. Since apache/arrow#51691, Arrow C++ keeps the input nullability of directly projected fields, so a build against Arrow C++ `main` returns `ID` as `not null` and the test fails. That is why the nightly Binaries job fails. This declares `ID` as `NULLABILITY_NULLABLE`, which matches the file. `named_table_users.json` also declares `ID` as required, but the named-table tests read a table supplied from Java with a nullable `ID`, and they pass against both versions, so it's unchanged. I ran `TestAceroSubstraitConsumer` with the `jni-macos-aarch_64` libraries from the RC workflow, which are built from Arrow C++ `main`. Before this change `testRunQueryLocalFiles` fails with the nightly's error; after it, all 11 tests pass. Arrow C++ 26.0.1 projects every field as nullable, so the result there doesn't change. Closes #1454. -- 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]
