Test for Hive compatibility across file formats

[findinpath]

Test whether the information written by Trino is correctly read by Hive across the following file formats:

ORC
PARQUET
RCBINARY
RCTEXT
SEQUENCEFILE
TEXTFILE
AVRO

This is the second attempt to merge these changes after #10309 (which has been reverted after finding the issue #10486)

[findepi]

cc @hashhar @kokosing

[hashhar]

Looks good to me. A question for my own understanding.

[findepi]

@findinpath the tests discovered a preexisting standing bug in the new Parquet writer -- #4608 (i reopened #6377 for that).

Please accomodate the current state of things to make the tests happy, so that we can merge the PR and prevent future regressions.

[findepi]

@findinpath it seems you did a rebase since the initial post, at which point it was a mere cherry pick.
What are the changes compared to #10309 that i should review?

[findinpath]

@findepi these are the changes that have appeared on the PR since the last review:

        if (!(isParquetStorageFormat && isParquetOptimizedWriterEnabled)) {
            // Hive expects `FIXED_LEN_BYTE_ARRAY` for decimal values irrespective of the Parquet specification which allows `INT32`, `INT64` for short precision decimal types
            columnDataList.add(new HiveCompatibilityColumnData("c_decimal_10_0", "decimal(10,0)", "346", new BigDecimal("346")));
            columnDataList.add(new HiveCompatibilityColumnData("c_decimal_10_2", "decimal(10,2)", "12345678.91", new BigDecimal("12345678.91")));
        }

if (!(isParquetStorageFormat && isParquetOptimizedWriterEnabled)) {
            // Hive expects `INT96` (deprecated on Parquet) for timestamp values
            ...
}

[findepi]

As a follow-up please add tests showing the timestamp and decimal-related Parquet problems.
They will be failing for now (io.trino.tempto.assertions.QueryAssert#assertQueryFailure)

[findinpath]

Created #10613 follow-up PR