SparkSQL is widely use to run ETL queries, why not support for it? or can we change https://github.com/sqlancer/sqlancer/pull/881 (for presto) to for SparkSQL?