Package org.apache.jorphan.math

Class StatCalculator<T extends Number & Comparable<? super T>>

java.lang.Object

org.apache.jorphan.math.StatCalculator<T>

Type Parameters:

T - type parameter for the calculator

Direct Known Subclasses:

StatCalculatorInteger, StatCalculatorLong

public abstract class StatCalculator<T extends Number & Comparable<? super T>>
extends Object

This class serves as a way to calculate the median, max, min etc. of a list of values. It is not threadsafe.

Constructor Summary

Constructors

Constructor	Description
StatCalculator(T zero, T min, T max)	This constructor is used to set up particular values for the generic class instance.

Method Summary

Modifier and Type	Method	Description
void	addAll(StatCalculator<T> calc)
void	addBytes(long newValue)
void	addValue(T val)	Add a single value (normally elapsed time)
void	addValue(T val, long sampleCount)	Update the calculator with the value for an aggregated sample.
void	clear()
protected abstract T	divide(T val, int n)
protected abstract T	divide(T val, long n)
long	getCount()
Map<Number,Number[]>	getDistribution()	Returns the distribution of the values in the list.
T	getMax()
double	getMean()
T	getMedian()
T	getMin()
T	getPercentPoint(double percent)	Get the value which %percent% of the values are less than.
T	getPercentPoint(float percent)	Get the value which %percent% of the values are less than.
double	getStandardDeviation()
double	getSum()
long	getTotalBytes()

Methods inherited from class java.lang.Object

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

Constructor Detail

StatCalculator

public StatCalculator(T zero,
                      T min,
                      T max)

This constructor is used to set up particular values for the generic class instance.

Parameters:

zero - - value to return for Median and PercentPoint if there are no values

min - - value to return for minimum if there are no values

max - - value to return for maximum if there are no values

Method Detail

clear

public void clear()

addBytes

public void addBytes(long newValue)

addAll

public void addAll(StatCalculator<T> calc)

getMedian

public T getMedian()

getTotalBytes

public long getTotalBytes()

getPercentPoint

public T getPercentPoint(float percent)

Get the value which %percent% of the values are less than. This works just like median (where median represents the 50% point). A typical desire is to see the 90% point - the value that 90% of the data points are below, the remaining 10% are above.

Parameters:

percent - number representing the wished percent (between 0 and 1.0)

Returns:

number of values less than the percentage

getPercentPoint

public T getPercentPoint(double percent)

Get the value which %percent% of the values are less than. This works just like median (where median represents the 50% point). A typical desire is to see the 90% point - the value that 90% of the data points are below, the remaining 10% are above.

Parameters:

percent - number representing the wished percent (between 0 and 1.0)

Returns:

the value which %percent% of the values are less than

getDistribution

public Map<Number,Number[]> getDistribution()

Returns the distribution of the values in the list.

Returns:

map containing either Integer or Long keys; entries are a Number array containing the key and the [Integer] count. TODO - why is the key value also stored in the entry array? See Bug 53825

getMean

public double getMean()

getStandardDeviation

public double getStandardDeviation()

getMin

public T getMin()

getMax

public T getMax()

getCount

public long getCount()

getSum

public double getSum()

divide

protected abstract T divide(T val,
                            int n)

divide

protected abstract T divide(T val,
                            long n)

addValue

public void addValue(T val,
                     long sampleCount)

Update the calculator with the value for an aggregated sample.

Parameters:

val - the aggregate value, normally the elapsed time

sampleCount - the number of samples contributing to the aggregate value

addValue

public void addValue(T val)

Add a single value (normally elapsed time)

Parameters:

val - the value to add, which should correspond with a single sample

See Also:

addValue(Number, long)