[FEA] Spark 3.1 orc nested predicate pushdown support

[tgravescs]

Describe the bug
Spark 3.1 added nested predicate pushdown support in ORC - apache/spark@7b6e1d5

This removed a function we were using:
E Caused by: java.lang.NoSuchMethodError: org.apache.spark.sql.execution.datasources.orc.OrcFiltersBase.isSearchableType$(Lorg/apache/spark/sql/execution/datasources/orc/OrcFiltersBase;Lorg/apache/spark/sql/types/DataType;)Z

We should go through and add the same support or fix it to use the new function.

Steps/Code to reproduce bug
Run spark 3.1.0 integration tests:
FAILED integration_tests/src/main/python/orc_test.py::test_input_meta - py4j....

Expected behavior
tests pass

Environment details (please complete the following information)

• Environment location: [Standalone, YARN, Kubernetes, Cloud(specify cloud provider)]
• Spark configuration settings related to the issue

Additional context
Add any other context about the problem here.

[tgravescs]

this got backported to 3.0.1 so we need to fix for 0.2 release

[tgravescs]

it looks like changes in 3.0.1 and 3.1.0 are also different so we would need a shim layer. Temporarily I'm going to put in a change to just copy the function that got moved to private:

• protected[sql] def isSearchableType(dataType: DataType) = dataType match {

• private def isSearchableType(dataType: DataType) = dataType match {

And then we can look at this in more detail to support it fully.

[tgravescs]

ok so leaving a P2 because we fixed the function and it works, this is only needed if we want to support nested predicate pushdown

[sameerz]

Depends on rapidsai/cudf#7640 and rapidsai/cudf#7830

[tgravescs]

our orc filter code was updated with #1982 and it looks like this should just work, so perhaps we just need to test it.

[jlowe]

Closing as a duplicate of #1481