[BUG] Non-nullable bools in a nullable struct fails

[kuhushukla]

Describe the bug
While working on a test for issue #11736. I see failures with below stack trace when booleans that have nullable=false are passed into structs with nullable=true .

Caused by: ai.rapids.cudf.CudfException: CUDF failure at: /home/jenkins/agent/workspace/jenkins-spark-rapids-jni-release-4-cuda12/thirdparty/cudf/cpp/src/io/orc/writer_impl.cu:586: Mismatch in metadata prescribed nullability and input column. Metadata for input column with nulls cannot prescribe nullability = false
E                   	at ai.rapids.cudf.Table.writeORCChunk(Native Method)
E                   	at ai.rapids.cudf.Table.access$1300(Table.java:41)
E                   	at ai.rapids.cudf.Table$ORCTableWriter.write(Table.java:2038)
E                   	at com.nvidia.spark.rapids.ColumnarOutputWriter.$anonfun$encodeAndBufferToHost$1(ColumnarOutputWriter.scala:205)
E                   	at com.nvidia.spark.rapids.ColumnarOutputWriter.$anonfun$encodeAndBufferToHost$1$adapted(ColumnarOutputWriter.scala:196)
E                   	at com.nvidia.spark.rapids.Arm$.withResource(Arm.scala:30)
E                   	at com.nvidia.spark.rapids.ColumnarOutputWriter.encodeAndBufferToHost(ColumnarOutputWriter.scala:196)
E                   	at com.nvidia.spark.rapids.ColumnarOutputWriter.$anonfun$bufferBatchAndClose$2(ColumnarOutputWriter.scala:180)
E                   	at com.nvidia.spark.rapids.ColumnarOutputWriter.$anonfun$bufferBatchAndClose$2$adapted(ColumnarOutputWriter.scala:179)

Steps/Code to reproduce bug
orc_write_test.py , pass BooleanGen(nullable=False) to StructGen and run the tests.

orc_write_basic_gens = [byte_gen, short_gen, int_gen, long_gen, float_gen, double_gen,
        string_gen, BooleanGen(nullable=False), DateGen(start=date(1590, 1, 1)),
        TimestampGen(start=datetime(1970, 1, 1, tzinfo=timezone.utc)) ] + \
        decimal_gens

orc_write_basic_struct_gen = StructGen([['child'+str(ind), sub_gen] for ind, sub_gen in enumerate(orc_write_basic_gens)])

Run command :

TEST_PARALLEL=10 ./run_pyspark_from_build.sh  -k "orc_write_test"

Expected behavior
Not error, fallback or ideally allow for Spark parity with nullables being reflected thru the metadata field.

Environment details (please complete the following information)

• Any, onprem in this case

[mattahrens]

Make sure we re-test this with latest changes related to boolean writes.

[revans2]

I extended these tests to other types and it looks like it fails for all of them on ORC.

[revans2]

This also fails for parquet in the same way...