[Java] Java JNI with S3 support

[asfimport]

I'm trying to use Arrow Java (https://mvnrepository.com/artifact/org.apache.arrow/arrow-dataset/8.0.0) to read a parquet file from S3. It fails due to the native library being compiled without S3 support:

java.lang.RuntimeException: Got S3 URI but Arrow compiled without S3 support
at org.apache.arrow.dataset.file.JniWrapper.makeFileSystemDatasetFactory(Native Method)
at org.apache.arrow.dataset.file.FileSystemDatasetFactory.createNative(FileSystemDatasetFactory.java:35)
at org.apache.arrow.dataset.file.FileSystemDatasetFactory.<init>(FileSystemDatasetFactory.java:31)

Reporter: Artavazd Balaian / @REASY
Assignee: Artavazd Balaian / @REASY

PRs and other links:

• GitHub Pull Request #13157

_{Note: This issue was originally created as ARROW-16584. Please see the migration documentation for further details.}

[asfimport]

Alessandro Molina / @amol-:
I think we also want to ensure the tests have a test to confirm that it works and we need to make sure it gets run on the verify-rc process of nightlies to ensure the package we build has S3 properly enabled.

[asfimport]

Artavazd Balaian / @REASY:
@amol- could you give some guidelines on how to do it? I can see in the case of C++ there are tests that use Minio (https://github.com/apache/arrow/blob/master/cpp/src/arrow/filesystem/s3fs_test.cc)

Should I just add a simple test that reads a parquet file from S3 (Minio) under https://github.com/apache/arrow/tree/master/java/dataset/src/test/java/org/apache/arrow/dataset ?  I don't see any use of docker in tests for Java, should I bring Minio as docker dependency or should I expect Minio executable to be always available?

[asfimport]

Kouhei Sutou / @kou:
Issue resolved by pull request 13157
#13157