GenSparkProcContext maintains information about the tasks and operators as we walk the operator tree to break them into SparkTasks.
GenSparkUtils is a collection of shared helper methods to produce SparkWork Cloned from GenTezUtils.
GenSparkWork separates the operator tree into spark tasks.
Walks the operator tree in DFS fashion.
SparkCompiler translates the operator plan into SparkTasks.
FileSinkProcessor handles addition of merge, move and stats tasks for filesinks.
This operator gets partition info from the upstream operators, and write them to HDFS.
ProcessAnalyzeTable sets up work for the several variants of analyze table (normal, no scan, partial scan.) The plan at this point will be a single table scan operator.
Data structure to keep track of SMBMapJoin operators during query compilation for Spark.
This processor triggers on SparkPartitionPruningSinkOperator.
Copyright © 2016 The Apache Software Foundation. All rights reserved.