[SPARK-19335][SPARK-38200][SQL] Add upserts for writing to JDBC using MERGE INTO with temp table

[EnricoMi]

What changes were proposed in this pull request?

Implements upsert mode for SaveMode.Append of the MsSql, Postgres, Derby, H2 and oracle JDBC source.

This uses MERGE INTO in combination with a temporary table. A batch of rows is inserted into the temporary table (rather than the target table) and merged into the target table with one MERGE INTO command per batch.

See #41518 for an alternative for databases not supporting MERGE INTO syntax.

Why are the changes needed?

The JDBC writer only supports either truncating the existing table or inserting. Duplicates, i.e. rows with identical values in the primary or unique index columns, cause an exception, permitting updating existing and inserting new rows.

Re-evaluating a partition due to executor loss will insert rows that have been inserted in an earlier attempt, which kills the entier Spark job.

Does this PR introduce any user-facing change?

This adds upsert and upsertKeyColumns options for SaveMode.Append of the JDBC source.

How was this patch tested?

Tests in JdbcSuite and integration suites.

[github-actions]

We're closing this PR because it hasn't been updated in a while. This isn't a judgement on the merit of the PR in any way. It's just a way of keeping the PR queue manageable.
If you'd like to revive this PR, please reopen it and ask a committer to remove the Stale tag!