Package org.apache.hadoop.hbase.metrics.impl

Class FastLongHistogram

java.lang.Object

org.apache.hadoop.hbase.metrics.impl.FastLongHistogram

@Private
@Evolving
public class FastLongHistogram
extends Object

FastLongHistogram is a thread-safe class that estimate distribution of data and computes the quantiles.

Nested Class Summary

Nested Classes

Modifier and Type	Class	Description
private static class	FastLongHistogram.Bins	Bins is a class containing a list of buckets(or bins) for estimation histogram of some data.

Field Summary

Fields

Modifier and Type	Field	Description
private FastLongHistogram.Bins	bins
static final int	DEFAULT_NBINS	Default number of bins.
static final double[]	DEFAULT_QUANTILES

Constructor Summary

Constructors

Constructor	Description
FastLongHistogram()	Constructor.
FastLongHistogram(int numOfBins)	Constructor.
FastLongHistogram(int numOfBins, long min, long max)	Constructor setting the bins assuming a uniform distribution within a range.

Method Summary

Modifier and Type	Method	Description
void	add(long value, long count)	Adds a value to the histogram.
long	getCount()
long	getMax()
long	getMean()
long	getMin()
long	getNumAtOrBelow(long value)
long[]	getQuantiles()
long[]	getQuantiles(double[] quantiles)	Computes the quantiles give the ratios.
Snapshot	snapshotAndReset()	Resets the histogram for new counting.

Methods inherited from class java.lang.Object

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

Field Details

DEFAULT_NBINS

public static final int DEFAULT_NBINS

Default number of bins.

See Also:

• Constant Field Values

DEFAULT_QUANTILES

public static final double[] DEFAULT_QUANTILES

bins

private volatile FastLongHistogram.Bins bins

Constructor Details

FastLongHistogram

public FastLongHistogram()

Constructor.

FastLongHistogram

public FastLongHistogram(int numOfBins)

Constructor.

Parameters:

numOfBins - the number of bins for the histogram. A larger value results in more precise results but with lower efficiency, and vice versus.

FastLongHistogram

public FastLongHistogram(int numOfBins,
 long min,
 long max)

Constructor setting the bins assuming a uniform distribution within a range.

Parameters:

numOfBins - the number of bins for the histogram. A larger value results in more precise results but with lower efficiency, and vice versus.

min - lower bound of the region, inclusive.

max - higher bound of the region, inclusive.

Method Details

add

public void add(long value,
 long count)

Adds a value to the histogram.

getQuantiles

public long[] getQuantiles(double[] quantiles)

Computes the quantiles give the ratios.

getQuantiles

public long[] getQuantiles()

getMin

public long getMin()

getMax

public long getMax()

getCount

public long getCount()

getMean

public long getMean()

getNumAtOrBelow

public long getNumAtOrBelow(long value)

snapshotAndReset

public Snapshot snapshotAndReset()

Resets the histogram for new counting.