native_iceberg_compat uses hard-coded config values

[parthchandra]

Describe the bug

In native_iceberg_compat, the initialization had codes the following configuration flags -

        conf.set("spark.sql.parquet.binaryAsString", "false");
        conf.set("spark.sql.parquet.int96AsTimestamp", "true");
        conf.set("spark.sql.caseSensitive", "false");
        conf.set("spark.sql.parquet.inferTimestampNTZ.enabled", "true");
        conf.set("spark.sql.legacy.parquet.nanosAsLong", "false");

These explicitly set the config to the default values for the configs.
We should apply the values specified by the end user instead and use defaults only if the value is not provided.

Steps to reproduce

No response

Expected behavior

No response

Additional context

No response

[mbutrovich]

Closed by #4363.