GH-1134: add COLUMN_DEF support to JDBC getColumnns()

[Verest]

What's Changed

Updates the FlightSqlColumnMetadata to contain the JDBC COLUMN_DEF field alongside getters/setters. Following exiting naming format.

Updates ArrowDatabaseMetadata to write to the pre-existing COLUMN_DEF vector as part of the getColumns() code paths.

Updates test(s) in ArrowDatabaseMetadataTest to test default value behavior.

Closes #1134.

[github-actions]

Thank you for opening a pull request!

Please label the PR with one or more of:

• bug-fix
• chore
• dependencies
• documentation
• enhancement

Also, add the 'breaking-change' label if appropriate.

See CONTRIBUTING.md for details.

[Verest]

I am murky on the specific format for COLUMN_DEF, or where that responsibility lies. The Oracle docs mention wrapping the underling value in single quotes if it is a string, and a few other things. I suspect this format is not followed everywhere though 🤔.

Since this is a generic JDBC driver, I am cautious to make assumptions about the incoming format of the COLUMN_DEF from the server. For now, I am tentatively expecting the server-side to handle the intended parsing of the COLUMN_DEF, leaving the arrow driver to read given strings as is.

Open to suggestions here.

Aside: this PR should be an enhancement, but I do not have permission to edit the labels.

[xborder]

Should this be discussed and voted like it was done for is_update? Maybe it is overkill but I think at least apache/arrow/FlightSql.proto should be in sync.

There are other places in FlightSql.proto that also mention metadata returned that probably would need to be updated

[xborder]

Have you thought if it would make sense in CommandGetXdbcTypeInfo?

[ennuite]

Should this be discussed and voted like it was done for is_update? Maybe it is overkill but I think at least apache/arrow/FlightSql.proto should be in sync.

There are other places in FlightSql.proto that also mention metadata returned that probably would need to be updated

Yes, the proto files in this repo should mirror the main one.

[jbonofre]

Thanks to have update the javadoc here.

I think it's worth to document COLUMN_DEF in the javadoc of CommandStatementQuery, CommandStatementSubstraitPlan, CommandPreparedStatementQuery. As we have REMARKS in these messages, we should have COLUMN_DEF.

[Verest]

Yes, I was uncertain about that - I can add them.

I did not add them originally as I do not think the COLUMN_DEF will be used in those code paths, but I don't think it hurts having it there either.

[jbonofre]

By the way, as this PR touches .proto, worth to at discuss on the mailing list.

[Verest]

Have you thought if it would make sense in CommandGetXdbcTypeInfo?

Off-hand I would say a defaultValue is N/A for that class as it appears to be generic to SQL types rather than specific columns - A default value is only relevant for a specific column of a table.

Maybe something along the lines of "supports_default_values" could be applicable there, but that would be out of scope for this PR. I am murky on the specifics of that class regardless.