org.datavec.api.transform.reduce

Class Reducer.Builder

java.lang.Object

org.datavec.api.transform.reduce.Reducer.Builder

Enclosing class:

Reducer

public static class Reducer.Builder
extends java.lang.Object

Constructor Summary

Constructors

Constructor and Description
Builder(ReduceOp defaultOp) Create a Reducer builder, and set the default column reduction operation.

Method Summary

Modifier and Type	Method and Description
Reducer	build()
Reducer.Builder	conditionalReduction(java.lang.String column, java.lang.String outputName, ReduceOp reduction, Condition condition) Conditional reduction: apply the reduce on a specified column, where the reduction occurs *only* on those examples where the condition returns true.
Reducer.Builder	countColumns(java.lang.String... columns) Reduce the specified columns by counting the number of values
Reducer.Builder	countUniqueColumns(java.lang.String... columns) Reduce the specified columns by counting the number of unique values
Reducer.Builder	customReduction(java.lang.String column, ColumnReduction columnReduction) Reduce the specified column using a custom column reduction functionality.
Reducer.Builder	keyColumns(java.lang.String... keyColumns) Specify the key columns.
Reducer.Builder	maxColumn(java.lang.String... columns) Reduce the specified columns by taking the maximum value
Reducer.Builder	meanColumns(java.lang.String... columns) Reduce the specified columns by taking the mean of the values
Reducer.Builder	minColumns(java.lang.String... columns) Reduce the specified columns by taking the minimum value
Reducer.Builder	rangeColumns(java.lang.String... columns) Reduce the specified columns by taking the range (max-min) of the values
Reducer.Builder	setIgnoreInvalid(java.lang.String... columns) When doing the reduction: set the specified columns to ignore any invalid values.
Reducer.Builder	stdevColumns(java.lang.String... columns) Reduce the specified columns by taking the standard deviation of the values
Reducer.Builder	sumColumns(java.lang.String... columns) Reduce the specified columns by taking the sum of values
Reducer.Builder	takeFirstColumns(java.lang.String... columns) Reduce the specified columns by taking the first value
Reducer.Builder	takeLastColumns(java.lang.String... columns) Reduce the specified columns by taking the last value

Methods inherited from class java.lang.Object

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

Constructor Detail

Builder

public Builder(ReduceOp defaultOp)

Create a Reducer builder, and set the default column reduction operation. For any columns that aren't specified explicitly, they will use the default reduction operation. If a column does have a reduction operation explicitly specified, then it will override the default specified here.

Parameters:

defaultOp - Default reduction operation to perform

Method Detail

keyColumns

public Reducer.Builder keyColumns(java.lang.String... keyColumns)

Specify the key columns. The idea here is to be able to create a (potentially compound) key out of multiple columns, using the toString representation of the values in these columns

Parameters:

keyColumns - Columns that will make up the key

Returns:

minColumns

public Reducer.Builder minColumns(java.lang.String... columns)

Reduce the specified columns by taking the minimum value

maxColumn

public Reducer.Builder maxColumn(java.lang.String... columns)

Reduce the specified columns by taking the maximum value

sumColumns

public Reducer.Builder sumColumns(java.lang.String... columns)

Reduce the specified columns by taking the sum of values

meanColumns

public Reducer.Builder meanColumns(java.lang.String... columns)

Reduce the specified columns by taking the mean of the values

stdevColumns

public Reducer.Builder stdevColumns(java.lang.String... columns)

Reduce the specified columns by taking the standard deviation of the values

countColumns

public Reducer.Builder countColumns(java.lang.String... columns)

Reduce the specified columns by counting the number of values

rangeColumns

public Reducer.Builder rangeColumns(java.lang.String... columns)

Reduce the specified columns by taking the range (max-min) of the values

countUniqueColumns

public Reducer.Builder countUniqueColumns(java.lang.String... columns)

Reduce the specified columns by counting the number of unique values

takeFirstColumns

public Reducer.Builder takeFirstColumns(java.lang.String... columns)

Reduce the specified columns by taking the first value

takeLastColumns

public Reducer.Builder takeLastColumns(java.lang.String... columns)

Reduce the specified columns by taking the last value

customReduction

public Reducer.Builder customReduction(java.lang.String column,
                                       ColumnReduction columnReduction)

Reduce the specified column using a custom column reduction functionality.

Parameters:

column - Column to execute the custom reduction functionality on

columnReduction - Column reduction to execute on that column

conditionalReduction

public Reducer.Builder conditionalReduction(java.lang.String column,
                                            java.lang.String outputName,
                                            ReduceOp reduction,
                                            Condition condition)

Conditional reduction: apply the reduce on a specified column, where the reduction occurs *only* on those examples where the condition returns true. Examples where the condition does not apply (returns false) are ignored/excluded.

Parameters:

column - Name of the column to execute the conditional reduction on

outputName - Name of the column, after the reduction has been executed

reduction - Reduction to execute

condition - Condition to use in the reductions

setIgnoreInvalid

public Reducer.Builder setIgnoreInvalid(java.lang.String... columns)

When doing the reduction: set the specified columns to ignore any invalid values. Invalid: defined as being not valid according to the ColumnMetaData: ColumnMetaData.isValid(Writable). For numerical columns, this typically means being unable to parse the Writable. For example, Writable.toLong() failing for a Long column. If the column has any restrictions (min/max values, regex for Strings etc) these will also be taken into account.

Parameters:

columns - Columns to set 'ignore invalid' for

build

public Reducer build()