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

Interface SampleEstimatorFactory

public interface SampleEstimatorFactory

Nested Class Summary

Nested Classes

Modifier and Type	Interface	Description
static class	SampleEstimatorFactory.EstimationType

Field Summary

Fields

Modifier and Type	Field	Description
static org.apache.commons.logging.Log	LOG

Method Summary

Static Methods

Modifier and Type	Method	Description
static int	distinctCount(int[] frequencies, int nRows, int sampleSize, SampleEstimatorFactory.EstimationType type)	Estimate a distinct number of values based on frequencies.
static int	distinctCount(int[] frequencies, int nRows, int sampleSize, SampleEstimatorFactory.EstimationType type, HashMap<Integer,Double> solveCache)	Estimate a distinct number of values based on frequencies.

Field Detail

LOG

static final org.apache.commons.logging.Log LOG

Method Detail

distinctCount

static int distinctCount(int[] frequencies,
                         int nRows,
                         int sampleSize,
                         SampleEstimatorFactory.EstimationType type)

Estimate a distinct number of values based on frequencies.

Parameters:

frequencies - A list of frequencies of unique values, NOTE all values contained should be larger than zero

nRows - The total number of rows to consider, NOTE should always be larger or equal to sum(frequencies)

sampleSize - The size of the sample, NOTE this should ideally be scaled to match the sum(frequencies) and should always be lower or equal to nRows

type - The type of estimator to use

Returns:

A estimated number of unique values

distinctCount

static int distinctCount(int[] frequencies,
                         int nRows,
                         int sampleSize,
                         SampleEstimatorFactory.EstimationType type,
                         HashMap<Integer,Double> solveCache)

Estimate a distinct number of values based on frequencies.

Parameters:

frequencies - A list of frequencies of unique values, NOTE all values contained should be larger than zero!

nRows - The total number of rows to consider, NOTE should always be larger or equal to sum(frequencies)

sampleSize - The size of the sample, NOTE this should ideally be scaled to match the sum(frequencies) and should always be lower or equal to nRows

type - The type of estimator to use

solveCache - A solve cache to avoid repeated calculations

Returns:

A estimated number of unique values