This new feature has been around and used/tests in many scenarios.
We should enable this feature by default.
Hive Native Reader reporting wrong results for timestamp column in hive generated parquet file
OOM with Hive native scan enabled on TPCH-100 parquet, query 05.q
Hive native reader is slow when the underlying parquet file has more row groups
Enhance HiveDrillNativeParquetScan and related classes to support multiple formats.
Enhance Scan costing to include factors other than row count
Parquet file with DATE logical type produces wrong results for simple SELECT
GitHub Pull Request #350
GitHub Pull Request #625