org.apache.flink.contrib.operatorstatistics

Class OperatorStatistics

java.lang.Object

org.apache.flink.contrib.operatorstatistics.OperatorStatistics

All Implemented Interfaces:

Serializable

public class OperatorStatistics
extends Object
implements Serializable

Data structure that encapsulates statistical information of data that has only been processed by one pass This statistical information is meant to help determine the distribution of the data that has been processed in an operator so that we can determine if it is necessary to repartition the data The statistics to be gathered are configurable and represented by a OperatorStatisticsConfig object. The information encapsulated in this class is min, max, a structure enabling estimation of count distinct and a structure holding the heavy hitters along with their frequency.

See Also:

Serialized Form

Constructor Summary

Constructors

Constructor and Description
OperatorStatistics(OperatorStatisticsConfig config)

Method Summary

Modifier and Type	Method and Description
OperatorStatistics	clone()
long	estimateCountDistinct()
Map<Object,Long>	getHeavyHitters()
Object	getMax()
Object	getMin()
void	merge(OperatorStatistics other)
void	process(Object tupleObject)
String	toString()

Methods inherited from class java.lang.Object

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

Constructor Detail

OperatorStatistics

public OperatorStatistics(OperatorStatisticsConfig config)

Method Detail

process

public void process(Object tupleObject)

merge

public void merge(OperatorStatistics other)
           throws RuntimeException

Throws:

RuntimeException

getMin

public Object getMin()

getMax

public Object getMax()

estimateCountDistinct

public long estimateCountDistinct()

getHeavyHitters

public Map<Object,Long> getHeavyHitters()

toString

public String toString()

Overrides:

toString in class Object

clone

public OperatorStatistics clone()

Overrides:

clone in class Object