Support PARTITIONED BY on Beam's SQL DDL

Partitioning by columns is a common optimization technique used in Hive and Spark to optimize queries performance.

Beam should support this feature to allow users that already have data stored following a partitioning schema to read and query it with Beam SQL.