Package org.apache.sysds.runtime.compress.estim.encoding

Class SparseEncoding

java.lang.Object

org.apache.sysds.runtime.compress.estim.encoding.AEncode

org.apache.sysds.runtime.compress.estim.encoding.SparseEncoding

All Implemented Interfaces:

IEncode

public class SparseEncoding
extends AEncode

A Encoding that contain a default value that is not encoded and every other value is encoded in the map. The logic is similar to the SDC column group.

Field Summary

Fields inherited from interface org.apache.sysds.runtime.compress.estim.encoding.IEncode

LOG

Method Summary

Modifier and Type	Method	Description
IEncode	combine(IEncode e)	Combine two encodings, note it should be guaranteed by the caller that the number of unique multiplied does not overflow Integer.
org.apache.commons.lang3.tuple.Pair<IEncode,Map<Integer,Integer>>	combineWithMap(IEncode e)	Combine two encodings without resizing the output.
boolean	equals(IEncode e)	Indicate if the given encoding is equivalent to this encoding
EstimationFactors	extractFacts(int nRows, double tupleSparsity, double matrixSparsity, CompressionSettings cs)	Extract the compression facts for this column group.
AMapToData	getMap()
int	getNumRows()
AOffset	getOffsets()
int	getUnique()	Get the number of unique values in this encoding
boolean	isDense()	Signify if the counts are including zero or without zero.
String	toString()

Methods inherited from class org.apache.sysds.runtime.compress.estim.encoding.AEncode

equals

Methods inherited from class java.lang.Object

getClass, hashCode, notify, notifyAll, wait, wait, wait

Method Detail

combine

public IEncode combine(IEncode e)

Description copied from interface: IEncode

Combine two encodings, note it should be guaranteed by the caller that the number of unique multiplied does not overflow Integer.

Parameters:

e - The other side to combine with

Returns:

The combined encoding

combineWithMap

public org.apache.commons.lang3.tuple.Pair<IEncode,Map<Integer,Integer>> combineWithMap(IEncode e)

Description copied from interface: IEncode

Combine two encodings without resizing the output. meaning the mapping of the indexes should be consistent with left hand side Dictionary indexes and right hand side indexes.

Parameters:

e - The other side to combine with

Returns:

The combined encoding

getUnique

public int getUnique()

Description copied from interface: IEncode

Get the number of unique values in this encoding

Returns:

The number of unique values.

extractFacts

public EstimationFactors extractFacts(int nRows,
                                      double tupleSparsity,
                                      double matrixSparsity,
                                      CompressionSettings cs)

Description copied from interface: IEncode

Extract the compression facts for this column group.

Parameters:

nRows - The total number of rows

tupleSparsity - The Sparsity of the unique tuples

matrixSparsity - The matrix sparsity

cs - The compression settings

Returns:

A EstimationFactors object

isDense

public boolean isDense()

Description copied from interface: IEncode

Signify if the counts are including zero or without zero.

Returns:

is dense

getOffsets

public AOffset getOffsets()

getMap

public AMapToData getMap()

getNumRows

public int getNumRows()

equals

public boolean equals(IEncode e)

Description copied from interface: IEncode

Indicate if the given encoding is equivalent to this encoding

Parameters:

e - The other encoding to be compared with this

Returns:

If the encoding is equivalent

toString

public String toString()

Overrides:

toString in class Object