Bulk Operation

FLINK-14133 add local aggregate to solve data skew for ROLLUP/CUBE case

FLINK-14133 Introduce data compression to blocking shuffle.

FLINK-14133 Introduce parallelism inference for HiveSource

FLINK-14133 extends max/min type in ColumnStats from Number to Comparable

FLINK-14133 Optimize BaseRowSerializer to use Projection instead of switch

FLINK-14133 Improve batch schedule check input consumable performance

FLINK-14133 Join condition could be simplified in logical phase

FLINK-14133 Optimize mapred.HadoopInputSplit to not serialize conf when split is not configurable

FLINK-14133 HiveTableSource should implement LimitableTableSource interface

FLINK-14133 Decide column nullability according to Hive constraints

FLINK-14133 blink planner should also fetch catalog statistics for permanent CatalogTableImpl

FLINK-14133 Introduce vectorized orc InputFormat for blink runtime

FLINK-14133 Introduce LimitableTableSource to optimize limit

FLINK-14133 HiveTableSource should implement ProjectableTableSource