fix: [df52] Fix/date32 int arithmetic

[andygrove]

Which issue does this PR close?

Closes #.

Rationale for this change

DataFusion 52's arrow-arith kernels only support Date32 +/- Interval types, not raw integers. When Spark sends Date32 + Int8/Int16/Int32 arithmetic, the planner now routes these operations to the Spark date_add/date_sub UDFs which handle integer types directly.

What changes are included in this PR?

How are these changes tested?

[comphead]

Thanks @andygrove LGTM would you care to describe what tests this PR addresses?

[andygrove]

Thanks @andygrove LGTM would you care to describe what tests this PR addresses?

Generated by AI:

The PR fixes Date32 +/- Int8/Int16/Int32 arithmetic which broke when upgrading to DataFusion 52, whose arrow-arith kernels only support Date32 +/- Interval types, not raw integers.

The following tests would fail without this fix:

CometExpressionSuite.scala

• "date_add with int scalars" (line ~301) — SELECT _20 + CAST(2 as TINYINT/SHORT/INT)
• "date_add with int arrays" (line ~335) — SELECT _20 + _2/_3/_4
• "date_sub with int scalars" (line ~349) — SELECT _20 - CAST(2 as TINYINT/SHORT/INT)
• "date_sub with int arrays" (line ~383) — SELECT _20 - _2/_3/_4

SQL file-driven tests (via CometSqlFileTestSuite)

• spark/src/test/resources/sql-tests/expressions/datetime/date_add.sql
• spark/src/test/resources/sql-tests/expressions/datetime/date_sub.sql

Fuzz tests (CometFuzzTemporalExpressions.scala)

• "date arithmetic - date_add"
• "date arithmetic - date_sub"

All of these tests perform Date32 + Int or Date32 - Int operations, which DataFusion 52's arrow-arith kernels no longer support. The fix routes those operations to the Spark date_add/date_sub UDFs
that handle integer types directly.