Enable sorted dynamic partitioning optimization for materialized views with custom data organization

After implementing HIVE-18842, we need to extend the optimizer to work with partitioned materialized views that are created with custom data organization, i.e., using CLUSTERED, DISTRIBUTED, or SORTED. Currently, optimization bails out when the materialized view is partitioned and either CLUSTERED, DISTRIBUTED, or SORTED.
In particular, we will need to combine the RS operator introduced by the translation of these clauses with the new RS needed to distribute and sort the data based on the dynamic partition values.