org.apache.beam.sdk.runners (Apache Beam SDK 0.2.0-incubating API)

Interface Summary
Interface	Description
PipelineRunnerRegistrar	`PipelineRunner` creators have the ability to automatically have their `PipelineRunner` registered with this SDK by creating a `ServiceLoader` entry and a concrete implementation of this interface.

Class Summary
Class	Description
AggregatorPipelineExtractor	Retrieves `Aggregators` at each `ParDo` and returns a `Map` of `Aggregator` to the `PTransforms` in which it is present.
AggregatorValues<T>	A collection of values associated with an `Aggregator`.
PipelineRunner<ResultT extends PipelineResult>	A `PipelineRunner` can execute, translate, or otherwise process a `Pipeline`.
TransformHierarchy	Captures information about a collection of transformations and their associated `PValue`s.
TransformTreeNode	Provides internal tracking of transform relationships with helper methods for initialization and ordered visitation.

Exception Summary
Exception Description

AggregatorRetrievalException
Signals that an exception has occurred while retrieving Aggregators.

Exception Summary
Exception	Description
AggregatorRetrievalException	Signals that an exception has occurred while retrieving `Aggregator`s.

Package org.apache.beam.sdk.runners Description

Defines runners for executing Pipelines in different modes, including org.apache.beam.sdk.runners.DirectRunner and org.apache.beam.sdk.runners.DataflowRunner.

org.apache.beam.sdk.runners.DirectRunner executes a Pipeline locally, without contacting the Dataflow service. org.apache.beam.sdk.runners.DataflowRunner submits a Pipeline to the Dataflow service, which executes it on Dataflow-managed Compute Engine instances. DataflowRunner returns as soon as the Pipeline has been submitted. Use org.apache.beam.sdk.runners.BlockingDataflowRunner to have execution updates printed to the console.

The runner is specified as part PipelineOptions.