ExpandExec Physical Operator¶

ExpandExec is a unary physical operator with support for Whole-Stage Java Code Generation.

Creating Instance¶

ExpandExec takes the following to be created:

• Projection Expressions (Seq[Seq[Expression]])
• Output Attributes
• Child physical operator

ExpandExec is created when:

• BasicOperators execution planning strategy is executed (to plan a Expand logical operator)

Performance Metrics¶

number of output rows¶

Generating Java Source Code for Consume Path¶

doConsume(
  ctx: CodegenContext,
  input: Seq[ExprCode],
  row: ExprCode): String

doConsume...FIXME

Generating Java Source Code for Produce Path¶

doProduce(
  ctx: CodegenContext): String

doProduce requests the child operator (that is supposed to be a CodegenSupport physical operator) to generate a Java source code for produce code path.

Executing Operator¶

doExecute(): RDD[InternalRow]

doExecute requests the child operator to execute (that creates a RDD[InternalRow]).

doExecute uses RDD.mapPartitions operator to apply a function to each partition of the RDD[InternalRow].

doExecute...FIXME

needCopyResult¶

needCopyResult: Boolean

needCopyResult is always enabled (true).

canPassThrough¶

ExpandExec canPassThrough in RemoveRedundantProjects physical optimization.

Demo¶