org.apache.hama.graph

Class AbstractAggregator<M extends org.apache.hadoop.io.Writable,VERTEX extends Vertex<?,?,M>>

java.lang.Object

org.apache.hama.graph.AbstractAggregator<M,VERTEX>

All Implemented Interfaces:

Aggregator<M,VERTEX>

Direct Known Subclasses:

AbsDiffAggregator, MaxAggregator, MinAggregator, SumAggregator

public abstract class AbstractAggregator<M extends org.apache.hadoop.io.Writable,VERTEX extends Vertex<?,?,M>>
extends Object
implements Aggregator<M,VERTEX>

Abstract base case of an aggregator. Also defines a sticky operation that aggregates the last value and the new one of a vertex.
For tracking cases it increments an internal counter on each call of aggregate.

Constructor Summary

Constructors

Constructor and Description
AbstractAggregator()

Method Summary

Methods

Modifier and Type	Method and Description
void	aggregate(VERTEX vertex, M value) Observes a value of a vertex after the compute method.
void	aggregate(VERTEX vertex, M oldValue, M newValue) Observes the old value of a vertex and the new value at the same time.
M	finalizeAggregation() Finalizes the aggregation on a master task.
org.apache.hadoop.io.IntWritable	getTimesAggregated()
M	getValue() Gets the value of the aggregator.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

AbstractAggregator

public AbstractAggregator()

Method Detail

aggregate

public void aggregate(VERTEX vertex,
             M value)

Observes a value of a vertex after the compute method. This is intended to be overriden by the user and is just an empty implementation in this class. Please make sure that you are null-checking vertex, since on a master task this will always be null.

Specified by:

aggregate in interface Aggregator<M extends org.apache.hadoop.io.Writable,VERTEX extends Vertex<?,?,M>>

aggregate

public void aggregate(VERTEX vertex,
             M oldValue,
             M newValue)

Observes the old value of a vertex and the new value at the same time. This is intended to be overridden by the user and is just an empty implementation in this class.Please make sure that you are null-checking vertex, since on a master task this will always be null.

finalizeAggregation

public M finalizeAggregation()

Finalizes the aggregation on a master task. This is intended to be overridden by the user and is just an empty implementation in this class (returns null).

getValue

public M getValue()

Gets the value of the aggregator. This is intended to be overridden by the user and is just an empty implementation in this class (returns null).

Specified by:

getValue in interface Aggregator<M extends org.apache.hadoop.io.Writable,VERTEX extends Vertex<?,?,M>>

getTimesAggregated

public org.apache.hadoop.io.IntWritable getTimesAggregated()