Support more unit tests from Spark 3.3/3.4

[zhouyuan]

Is your feature request related to a problem or challenge? Please describe what you are trying to do.
Gluten has ported the major part SQL based unit tests from Spark 3.2 branch, however upstream Spark has added more unit tests in 3.3 and 3.4 branch. Gluten does not check for those new tests

Describe the solution you'd like
Gluten should check for those new tests from Spark 3.3/3.4

The existing unit tests in Spark 3.2 should be refactored as common unit tests.

• In Spark 3.2 gluten should check those common unit tests
• In Spark 3.3 gluten should check those common unit tests and the new tests in 3.3

Describe alternatives you've considered
N/A

Additional context
N/a

CC: @zzcclp

[zhouyuan]

Here's the list of different tests in Spark 3.3 vs. Spark 3.2
https://gist.github.com/zhouyuan/8c41cb1b579b3ca5bb5879ff7260c139
So there are two kinds of changes:

• newly added tests -> put the new code on spark 3.3
• modified tests over Spark 3.2 due to API change 1) for removed unit tests, disbale them in common, enabled in spark 3.2 2) for modified tests, refactor them to move part into common UT, part into spark3.2, and part into spark3.3

[rui-mo]

Exclude should be able to resolve a persistent view from Spark 33 test in #1996 due to below issue.

[zhouyuan]

Below tests need to be fixed for Spark 3.3

execution/adaptive/GlutenAdaptiveQueryExecSuite.scala
execution/benchmarks/ParquetReadBenchmark.scala
execution/datasources/json/GlutenJsonSuite.scala

```
- Various partition value types
- Various inferred partition value types
- Various partition value types
- Various inferred partition value types
- SPARK-32908: maximum target error in percentile_approx

- SPARK-36825, SPARK-36854: year-month/day-time intervals written and read as INT32/INT64
- support batch reads for schema
- SPARK-36182: read TimestampNTZ as TimestampLTZ
- SPARK-36797: Union should resolve nested columns as top-level columns
- SPARK-37371: UnionExec should support columnar if all children support columnar
- SPARK-36280: Remove redundant aliases after RewritePredicateSubquery
- SPARK-36182: read TimestampNTZ as TimestampLTZ
- SPARK-39833: pushed filters with project without filter columns
- SPARK-36825, SPARK-36854: year-month/day-time intervals written and read as INT32/INT64
- support batch reads for schema
- SPARK-36794: Ignore duplicated key when building relation for semi/anti hash join

[PengleiShi]

Spark 3.4.1 has been a stable release, will gluten support 3.4 in the near future?

[zhouyuan]

closing with new issues opened