org.apache.sysds.runtime.compress.estim

Class CompressedSizeInfoColGroup

java.lang.Object

org.apache.sysds.runtime.compress.estim.CompressedSizeInfoColGroup

public class CompressedSizeInfoColGroup
extends Object

Information collected about a specific ColGroup's compression size.

Constructor Summary

Constructors

Constructor and Description
CompressedSizeInfoColGroup(EstimationFactors facts, Set<AColGroup.CompressionType> validCompressionTypes, ABitmap map) Main Constructor for accurate estimates.
CompressedSizeInfoColGroup(int[] columns, int numVals, int numRows) Join columns without analyzing the content.

Method Summary

Modifier and Type	Method and Description
static CompressedSizeInfoColGroup	addConstGroup(int[] columns, CompressedSizeInfoColGroup oneSide, Set<AColGroup.CompressionType> validCompressionTypes) This method adds a column group without having to analyze.
boolean	containsZeros()
boolean	equals(Object that)
Map<AColGroup.CompressionType,Long>	getAllCompressionSizes()
AColGroup.CompressionType	getBestCompressionType()
AColGroup.CompressionType	getBestCompressionType(CompressionSettings cs)
double	getCardinalityRatio()
int[]	getColumns()
long	getCompressionSize(AColGroup.CompressionType ct)
AMapToData	getMap()
long	getMinSize()
double	getMostCommonFraction()
int	getNumOffs() Number of offsets, or number of non zero values.
int	getNumVals() Note cardinality is the same as number of distinct values.
double	getTupleSparsity()
int	hashCode()
String	toString()

Methods inherited from class java.lang.Object

getClass, notify, notifyAll, wait, wait, wait

Constructor Detail

CompressedSizeInfoColGroup

public CompressedSizeInfoColGroup(int[] columns,
                                  int numVals,
                                  int numRows)

Join columns without analyzing the content. This only specify the compression ratio if encoded in DDC since this is trivially calculated. The number of tuples contained can be set to the upper theoretical bound of two groups by multiplying the number of distinct tuple of each colGroups with each other. SHOULD NOT BE USED FOR AN ACCURATE ESTIMATE OF SIZE

Parameters:

columns - The columns combined

numVals - The number of distinct value tuples contained

numRows - The number of rows.

CompressedSizeInfoColGroup

public CompressedSizeInfoColGroup(EstimationFactors facts,
                                  Set<AColGroup.CompressionType> validCompressionTypes,
                                  ABitmap map)

Main Constructor for accurate estimates.

Parameters:

facts - The facts extracted from a number of columns, based on the estimateFactors

validCompressionTypes - The list of valid compression types, allowed to be performed.

map - The map of the distinct values contained in this column group.

Method Detail

addConstGroup

public static CompressedSizeInfoColGroup addConstGroup(int[] columns,
                                                       CompressedSizeInfoColGroup oneSide,
                                                       Set<AColGroup.CompressionType> validCompressionTypes)

This method adds a column group without having to analyze. This is because the columns added are constant groups. NOTE THIS IS ONLY VALID IF THE COLUMN ADDED IS EMPTY OR CONSTANT!

Parameters:

columns - The columns of the colgroups together

oneSide - One of the sides, this may contain something, but the other side (not part of the argument) should not.

validCompressionTypes - The List of valid compression techniques to use

Returns:

A Combined estimate of the column group.

getCompressionSize

public long getCompressionSize(AColGroup.CompressionType ct)

getBestCompressionType

public AColGroup.CompressionType getBestCompressionType(CompressionSettings cs)

getBestCompressionType

public AColGroup.CompressionType getBestCompressionType()

getAllCompressionSizes

public Map<AColGroup.CompressionType,Long> getAllCompressionSizes()

getMinSize

public long getMinSize()

getNumVals

public int getNumVals()

Note cardinality is the same as number of distinct values.

Returns:

cardinality or number of distinct values.

getNumOffs

public int getNumOffs()

Number of offsets, or number of non zero values.

Returns:

Number of non zeros or number of values.

getColumns

public int[] getColumns()

getCardinalityRatio

public double getCardinalityRatio()

getMostCommonFraction

public double getMostCommonFraction()

getTupleSparsity

public double getTupleSparsity()

getMap

public AMapToData getMap()

containsZeros

public boolean containsZeros()

hashCode

public int hashCode()

Overrides:

hashCode in class Object

equals

public boolean equals(Object that)

Overrides:

equals in class Object

toString

public String toString()

Overrides:

toString in class Object