[SQL][TESTS][SPARK-52918] Batch JDBC database statements in JDBC suites

[alekjarmov]

What changes were proposed in this pull request?

To modify the before all to reduce the amount of roundtrips with the database. Per my benchmarks this led to decreased time in beforeAll from ~850ms to ~690ms in JDBCV2Suite and 3s to 1.8s in JDBCSuite. It also clears tech debt where people in the future won't unknowingly add more roundtrips than needed.

Why are the changes needed?

Improve test performance, it is not drastic improvement when running whole suite but is great when running just a single test.

Does this PR introduce any user-facing change?

No

How was this patch tested?

Test-only change.

Was this patch authored or co-authored using generative AI tooling?

Generated-by: Claiude Sonnet

[PetarVasiljevic-DB]

@alekjarmov I am interest in performance when we do addBatch, do you maybe know the numbers?

The diff would be minimal and it would be cleaner to read.

[alekjarmov]

@PetarVasiljevic-DB It should be around the same as we would have the same amount of network roundtrips which is where the most time is wasted, just the rewrite/diff would be more different as they use some different syntax with question marks.

[PetarVasiljevic-DB]

@PetarVasiljevic-DB It should be around the same as we would have the same amount of network roundtrips which is where the most time is wasted, just the rewrite/diff would be more different as they use some different syntax with question marks.

You wouldn't use question marks. You would create statement at the beginning and then instead of executing the statement, you just call addBatch instead of prepareStatement: https://www.baeldung.com/jdbc-batch-processing#1-batch-processing-using-statement

[cloud-fan]

in this suite, CREATE TABLE and INSERT are in one batch. Why can't we do the same for JDBCSuite.scala?

[alekjarmov]

Did it in JDBCSuite as well.