com.yahoo.bullet.querying.aggregations.sketches

Class QuantileSketch

java.lang.Object

com.yahoo.bullet.querying.aggregations.sketches.Sketch

com.yahoo.bullet.querying.aggregations.sketches.DualSketch

com.yahoo.bullet.querying.aggregations.sketches.QuantileSketch

public class QuantileSketch
extends DualSketch

Wraps operations for working with a DoublesSketch - Quantile Sketch.

Field Summary

Fields

Modifier and Type	Field and Description
static String	COUNT_FIELD
static String	END_EXCLUSIVE
static String	INFINITY
static String	NEGATIVE_INFINITY
static String	NEGATIVE_INFINITY_START
static String	POSITIVE_INFINITY
static String	POSITIVE_INFINITY_END
static String	PROBABILITY_FIELD
static String	QUANTILE_FIELD
static double	QUANTILE_MAX
static double	QUANTILE_MIN
static String	RANGE_FIELD
static String	SEPARATOR
static String	START_EXCLUSIVE
static String	START_INCLUSIVE
static String	VALUE_FIELD

Fields inherited from class com.yahoo.bullet.querying.aggregations.sketches.Sketch

provider

Constructor Summary

Constructors

Constructor and Description
QuantileSketch(int k, DistributionType type, double[] points, com.yahoo.bullet.record.BulletRecordProvider provider) Creates a quantile sketch with the given number of entries getting results with the given points.
QuantileSketch(int k, int rounding, DistributionType type, int numberOfPoints, com.yahoo.bullet.record.BulletRecordProvider provider) Creates a quantile sketch with the given number of entries generating results with the number of points (evenly-spaced).

Method Summary

Modifier and Type	Method and Description
protected Map<String,Object>	addMetadata(Map<String,String> conceptKeys) Adds the common metadata for this Sketch to Map.
protected String	getFamily() Returns a String representing the family of this sketch.
List<com.yahoo.bullet.record.BulletRecord>	getRecords() Collects the data presented to the sketch so far and returns the List of BulletRecord representation of the resulting data in the sketch.
Clip	getResult(String metaKey, Map<String,String> conceptKeys) Gets the result from the data presented to the sketch as a Clip.
protected Integer	getSize() Returns the size of the Sketch in bytes.
protected Boolean	isEstimationMode() Returns whether this sketch was in estimation mode or not.
protected void	mergeBothSketches() Merge and reset both the update sketch and the union sketch into the result.
protected void	mergeUnionSketch() Merge and reset just the union sketch into the result.
protected void	mergeUpdateSketch() Merge and reset just the update sketch into the result.
void	reset() Resets the Sketch to the original state.
byte[]	serialize() Serializes the sketch.
void	union(byte[] serialized) Union a sketch serialized using Sketch.serialize() into this.
protected boolean	unionedExistingResults() Merge the existing result into the union sketch if needed and return a boolean if the union was done.
void	update(double data) Updates the sketch with a double.

Methods inherited from class com.yahoo.bullet.querying.aggregations.sketches.DualSketch

merge, union, update

Methods inherited from class com.yahoo.bullet.querying.aggregations.sketches.Sketch

getMetadata

Methods inherited from class java.lang.Object

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

Field Detail

QUANTILE_MIN

public static final double QUANTILE_MIN

See Also:

Constant Field Values

QUANTILE_MAX

public static final double QUANTILE_MAX

See Also:

Constant Field Values

QUANTILE_FIELD

public static final String QUANTILE_FIELD

See Also:

Constant Field Values

VALUE_FIELD

public static final String VALUE_FIELD

See Also:

Constant Field Values

PROBABILITY_FIELD

public static final String PROBABILITY_FIELD

See Also:

Constant Field Values

COUNT_FIELD

public static final String COUNT_FIELD

See Also:

Constant Field Values

RANGE_FIELD

public static final String RANGE_FIELD

See Also:

Constant Field Values

START_INCLUSIVE

public static final String START_INCLUSIVE

See Also:

Constant Field Values

START_EXCLUSIVE

public static final String START_EXCLUSIVE

See Also:

Constant Field Values

END_EXCLUSIVE

public static final String END_EXCLUSIVE

See Also:

Constant Field Values

SEPARATOR

public static final String SEPARATOR

See Also:

Constant Field Values

INFINITY

public static final String INFINITY

See Also:

Constant Field Values

POSITIVE_INFINITY

public static final String POSITIVE_INFINITY

See Also:

Constant Field Values

NEGATIVE_INFINITY

public static final String NEGATIVE_INFINITY

See Also:

Constant Field Values

NEGATIVE_INFINITY_START

public static final String NEGATIVE_INFINITY_START

See Also:

Constant Field Values

POSITIVE_INFINITY_END

public static final String POSITIVE_INFINITY_END

See Also:

Constant Field Values

Constructor Detail

QuantileSketch

public QuantileSketch(int k,
                      DistributionType type,
                      double[] points,
                      com.yahoo.bullet.record.BulletRecordProvider provider)

Creates a quantile sketch with the given number of entries getting results with the given points.

Parameters:

k - A number representative of the size of the sketch.

type - A DistributionType that determines what the points mean.

points - An array of points to get the quantiles, PMF and/or CDF for.

provider - A BulletRecordProvider to generate BulletRecords.

QuantileSketch

public QuantileSketch(int k,
                      int rounding,
                      DistributionType type,
                      int numberOfPoints,
                      com.yahoo.bullet.record.BulletRecordProvider provider)

Creates a quantile sketch with the given number of entries generating results with the number of points (evenly-spaced).

Parameters:

k - A number representative of the size of the sketch.

rounding - A number representing how many max decimal places points should have.

type - A DistributionType that determines what the points mean.

numberOfPoints - A positive number of evenly spaced points in the range for the type to get the data for.

provider - A BulletRecordProvider to generate BulletRecords.

Method Detail

update

public void update(double data)

Updates the sketch with a double.

Parameters:

data - A double to insert into the sketch.

union

public void union(byte[] serialized)

Description copied from class: Sketch

Union a sketch serialized using Sketch.serialize() into this.

Specified by:

union in class Sketch

Parameters:

serialized - A sketch serialized using the serialize method.

serialize

public byte[] serialize()

Description copied from class: Sketch

Serializes the sketch.

Specified by:

serialize in class Sketch

Returns:

A byte[] representing the serialized sketch.

getRecords

public List<com.yahoo.bullet.record.BulletRecord> getRecords()

Description copied from class: Sketch

Collects the data presented to the sketch so far and returns the List of BulletRecord representation of the resulting data in the sketch. See Sketch.getResult(String, Map) for getting the result including the metadata and see Sketch.getMetadata(String, Map) for getting only the metadata.

Specified by:

getRecords in class Sketch

Returns:

A list of resulting records representing the result in the sketch.

getResult

public Clip getResult(String metaKey,
                      Map<String,String> conceptKeys)

Description copied from class: Sketch

Gets the result from the data presented to the sketch as a Clip. Also adds Meta if asked for.

Overrides:

getResult in class Sketch

Parameters:

metaKey - If set to a non-null value, Sketch metadata will be added to the result.

conceptKeys - If provided, these Meta.Concept names will be added to the metadata.

Returns:

A Clip of the results.

reset

public void reset()

Description copied from class: DualSketch

Resets the Sketch to the original state. The old results are lost. You should call this after you call Sketch.serialize() or Sketch.getResult(String, Map) if you want to add more data to the sketch. Call this after your reset operations to reset the metadata relating to storing two sketches. You should reset or remove all your sketches here.

Overrides:

reset in class DualSketch

mergeBothSketches

protected void mergeBothSketches()

Description copied from class: DualSketch

Merge and reset both the update sketch and the union sketch into the result.

Specified by:

mergeBothSketches in class DualSketch

mergeUpdateSketch

protected void mergeUpdateSketch()

Description copied from class: DualSketch

Merge and reset just the update sketch into the result.

Specified by:

mergeUpdateSketch in class DualSketch

mergeUnionSketch

protected void mergeUnionSketch()

Description copied from class: DualSketch

Merge and reset just the union sketch into the result.

Specified by:

mergeUnionSketch in class DualSketch

unionedExistingResults

protected boolean unionedExistingResults()

Description copied from class: DualSketch

Merge the existing result into the union sketch if needed and return a boolean if the union was done.

Specified by:

unionedExistingResults in class DualSketch

Returns:

A boolean denoting whether the union was done.

addMetadata

protected Map<String,Object> addMetadata(Map<String,String> conceptKeys)

Description copied from class: Sketch

Adds the common metadata for this Sketch to Map.

Overrides:

addMetadata in class Sketch

Parameters:

conceptKeys - The Map of Meta.Concept names to their keys.

Returns:

The created Map of sketch metadata.

isEstimationMode

protected Boolean isEstimationMode()

Description copied from class: Sketch

Returns whether this sketch was in estimation mode or not.

Specified by:

isEstimationMode in class Sketch

Returns:

A Boolean denoting whether this sketch was estimating.

getFamily

protected String getFamily()

Description copied from class: Sketch

Returns a String representing the family of this sketch.

Specified by:

getFamily in class Sketch

Returns:

The String family of this sketch.

getSize

protected Integer getSize()

Description copied from class: Sketch

Returns the size of the Sketch in bytes.

Specified by:

getSize in class Sketch

Returns:

An Integer representing the size of the sketch.