Package org.apache.sysds.runtime.data

Class SparseRowVector

java.lang.Object

org.apache.sysds.runtime.data.SparseRow

org.apache.sysds.runtime.data.SparseRowVector

All Implemented Interfaces:

Serializable

public final class SparseRowVector
extends SparseRow

A sparse row vector that is able to grow dynamically as values are appended to it.

See Also:

Serialized Form

Field Summary

Fields

Modifier and Type	Field	Description
static int	initialCapacity	Initial capacity of any created sparse row

Constructor Summary

Constructors

Constructor	Description
SparseRowVector()
SparseRowVector(double[] v)	Sparse row vector constructor that take a dense array, and allocate sparsely by ignoring zero values
SparseRowVector(double[] v, int[] i)
SparseRowVector(double[] v, int[] i, int nnz)
SparseRowVector(int capacity)
SparseRowVector(int nnz, double[] v, int vlen)
SparseRowVector(int estnnz, int maxnnz)
SparseRowVector(SparseRow that)

Method Summary

Modifier and Type	Method	Description
boolean	add(int col, double v)	Add a value to a specified column with awareness of potential insertions.
SparseRowVector	append(int col, double v)	Appends a value to the end of the sparse row.
int	capacity()
void	compact()	In-place compaction of non-zero-entries; removes zero entries and shifts non-zero entries to the left if necessary.
void	compact(double eps)	In-place compaction of values over eps away from zero; and shifts non-zero entries to the left if necessary.
SparseRow	copy(boolean deep)	Make a copy of this row.
void	copy(SparseRow that)
void	deleteIndexRange(int lowerCol, int upperCol)
double	get(int col)	Gets the value of a specified column.
int	getIndex(int col)
int[]	indexes()	Get the index array of non-zero entries, co-aligned with the array of values.
boolean	isEmpty()	Indicates if the sparse row is empty, i.e., if is has size zero.
void	reset(int estnns, int maxnns)	Resets the sparse row to empty, after this call size and isEmpty are guaranteed to return 0 and true, respectively.
int	searchIndexesFirstGT(int col)
int	searchIndexesFirstGTE(int col)	Get first index greater than equal column index.
int	searchIndexesFirstLTE(int col)
boolean	set(int col, double v)	Sets the value of a specified column with awareness of potential overwrites or deletes (set to value zero).
void	setAtPos(int pos, int col, double v)
void	setEstimatedNzs(int estnnz)
void	setIndexes(int[] i)
void	setIndexRange(int cl, int cu, double[] v, int[] vix, int vpos, int vlen)
void	setIndexRange(int cl, int cu, double[] v, int vix, int vlen)
void	setSize(int newsize)
void	setValues(double[] d)
int	size()	Get the number of non-zero values of the sparse row.
void	sort()	In-place sort of column-index value pairs in order to allow binary search after constant-time append was used for reading unordered sparse rows.
double[]	values()	Get the value array of non-zero entries, co-aligned with the array of indexes.

Methods inherited from class org.apache.sysds.runtime.data.SparseRow

toString

Methods inherited from class java.lang.Object

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

Field Detail

initialCapacity

public static final int initialCapacity

Initial capacity of any created sparse row

WARNING: be aware that this affects the core memory estimates (incl. implicit assumptions)!

See Also:

Constant Field Values

Constructor Detail

SparseRowVector

public SparseRowVector()

SparseRowVector

public SparseRowVector(int capacity)

SparseRowVector

public SparseRowVector(int nnz,
                       double[] v,
                       int vlen)

SparseRowVector

public SparseRowVector(double[] v,
                       int[] i)

SparseRowVector

public SparseRowVector(double[] v,
                       int[] i,
                       int nnz)

SparseRowVector

public SparseRowVector(double[] v)

Sparse row vector constructor that take a dense array, and allocate sparsely by ignoring zero values

Parameters:

v - The dense row

SparseRowVector

public SparseRowVector(int estnnz,
                       int maxnnz)

SparseRowVector

public SparseRowVector(SparseRow that)

Method Detail

size

public int size()

Description copied from class: SparseRow

Get the number of non-zero values of the sparse row.

Specified by:

size in class SparseRow

Returns:

number of non-zeros

setSize

public void setSize(int newsize)

isEmpty

public final boolean isEmpty()

Description copied from class: SparseRow

Indicates if the sparse row is empty, i.e., if is has size zero.

Specified by:

isEmpty in class SparseRow

Returns:

true if empty

values

public double[] values()

Description copied from class: SparseRow

Get the value array of non-zero entries, co-aligned with the array of indexes.

Specified by:

values in class SparseRow

Returns:

array of values

indexes

public int[] indexes()

Description copied from class: SparseRow

Get the index array of non-zero entries, co-aligned with the array of values.

Specified by:

indexes in class SparseRow

Returns:

array of indexes

setValues

public void setValues(double[] d)

setIndexes

public void setIndexes(int[] i)

capacity

public int capacity()

copy

public void copy(SparseRow that)

reset

public void reset(int estnns,
                  int maxnns)

Description copied from class: SparseRow

Resets the sparse row to empty, after this call size and isEmpty are guaranteed to return 0 and true, respectively.

Specified by:

reset in class SparseRow

Parameters:

estnns - estimated number of non-zeros

maxnns - maximum number of non-zeros, e.g., number of columns

setEstimatedNzs

public void setEstimatedNzs(int estnnz)

set

public boolean set(int col,
                   double v)

Description copied from class: SparseRow

Sets the value of a specified column with awareness of potential overwrites or deletes (set to value zero).

Specified by:

set in class SparseRow

Parameters:

col - column index, zero-based

v - value

Returns:

true if the size of the sparse row changed

setAtPos

public void setAtPos(int pos,
                     int col,
                     double v)

add

public boolean add(int col,
                   double v)

Description copied from class: SparseRow

Add a value to a specified column with awareness of potential insertions.

Specified by:

add in class SparseRow

Parameters:

col - column index, zero-based

v - value

Returns:

true if the size of the sparse row changed

append

public SparseRowVector append(int col,
                              double v)

Description copied from class: SparseRow

Appends a value to the end of the sparse row.

Specified by:

append in class SparseRow

Parameters:

col - column index, zero-based

v - value

Returns:

the row with an appended element

get

public double get(int col)

Description copied from class: SparseRow

Gets the value of a specified column. If the column index does not exist in the sparse row, this call returns zero.

Specified by:

get in class SparseRow

Parameters:

col - column index, zero-based

Returns:

value

getIndex

public int getIndex(int col)

searchIndexesFirstLTE

public int searchIndexesFirstLTE(int col)

searchIndexesFirstGTE

public int searchIndexesFirstGTE(int col)

Description copied from class: SparseRow

Get first index greater than equal column index.

Specified by:

searchIndexesFirstGTE in class SparseRow

Parameters:

col - column to be greater than

Returns:

index

searchIndexesFirstGT

public int searchIndexesFirstGT(int col)

Specified by:

searchIndexesFirstGT in class SparseRow

deleteIndexRange

public void deleteIndexRange(int lowerCol,
                             int upperCol)

setIndexRange

public void setIndexRange(int cl,
                          int cu,
                          double[] v,
                          int vix,
                          int vlen)

setIndexRange

public void setIndexRange(int cl,
                          int cu,
                          double[] v,
                          int[] vix,
                          int vpos,
                          int vlen)

sort

public void sort()

Description copied from class: SparseRow

In-place sort of column-index value pairs in order to allow binary search after constant-time append was used for reading unordered sparse rows. We first check if already sorted and subsequently sort if necessary in order to get O(n) best case. Note: In-place sort is necessary in order to guarantee the memory estimate for operations that implicitly read that data set.

Specified by:

sort in class SparseRow

compact

public void compact()

Description copied from class: SparseRow

In-place compaction of non-zero-entries; removes zero entries and shifts non-zero entries to the left if necessary.

Specified by:

compact in class SparseRow

compact

public void compact(double eps)

Description copied from class: SparseRow

In-place compaction of values over eps away from zero; and shifts non-zero entries to the left if necessary.

Specified by:

compact in class SparseRow

Parameters:

eps - epsilon value

copy

public SparseRow copy(boolean deep)

Description copied from class: SparseRow

Make a copy of this row.

Specified by:

copy in class SparseRow

Parameters:

deep - if the copy should be deep

Returns:

A copy