Fallback to Spark in native_datafusion/native_iceberg_compat if encryption is enabled

[parthchandra]

Describe the bug

Not entirely sure if this will work but per the docs -

// Activate Parquet encryption, driven by Hadoop properties
sc.hadoopConfiguration().set("parquet.crypto.factory.class" ,
   "org.apache.parquet.crypto.keytools.PropertiesDrivenCryptoFactory");

So maybe we can fallback to Spark if parquet.crypto.factory.class is defined.

Steps to reproduce

No response

Expected behavior

No response

Additional context

No response