org.apache.sysds.runtime.compress.colgroup

Class AColGroup

java.lang.Object

org.apache.sysds.runtime.compress.colgroup.AColGroup

All Implemented Interfaces:

Serializable

Direct Known Subclasses:

ColGroupCompressed, ColGroupUncompressed

public abstract class AColGroup
extends Object
implements Serializable

Abstract Class that is the lowest class type for the Compression framework. AColGroup store information about a number of columns.

See Also:

Serialized Form

Nested Class Summary

Nested Classes

Modifier and Type	Class and Description
static class	AColGroup.CompressionType Public super types of compression ColGroups supported

Method Summary

Modifier and Type	Method and Description
abstract AColGroup	binaryRowOp(BinaryOperator op, double[] v, boolean sparseSafe, boolean left) Perform a binary row operation.
static double[]	colSum(List<AColGroup> groups, double[] res, int nRows) Compute the column sum of the given list of groups
abstract void	computeColSums(double[] c, int nRows) Compute the column sum
abstract boolean	containsValue(double pattern) Detect if the column group contains a specific value.
abstract AColGroup	copy() Get a copy of this column group.
abstract void	countNonZerosPerRow(int[] rnnz, int rl, int ru) Count the number of non-zeros per row
void	decompressToBlock(MatrixBlock target, int rl, int ru) Decompress the contents of the column group without modifying nonZeroCount.
abstract void	decompressToBlock(MatrixBlock target, int rl, int ru, int offT) Decompress the contents of the column group without counting non zeros
long	estimateInMemorySize() Get the upper bound estimate of in memory allocation for the column group.
abstract double	get(int r, int c) Get the value at a global row/column position.
int[]	getColIndices() Obtain the offsets of the columns in the matrix block that make up the group
abstract AColGroup.CompressionType	getCompType() Obtain the compression type.
long	getExactSizeOnDisk() Returns the exact serialized size of column group.
abstract double	getMax() Short hand method for getting maximum value contained in this column group.
abstract double	getMin() Short hand method for getting minimum value contained in this column group.
abstract long	getNumberNonZeros(int nRows) Get the number of nonZeros contained in this column group.
int	getNumCols() Obtain the number of columns in this column group.
abstract int	getNumValues() Obtain number of distinct tuples in contained sets of values associated with this column group.
abstract double[]	getValues() Get all the values in the colGroup.
abstract MatrixBlock	getValuesAsBlock() Returns the ColGroup as a MatrixBlock.
abstract boolean	isDense() Is dense, signals that the entire column group is allocated an processed.
abstract boolean	isLossy() Is Lossy
abstract void	leftMultByAColGroup(AColGroup lhs, MatrixBlock result) Left side matrix multiplication with a column group that is transposed.
void	leftMultByMatrix(MatrixBlock matrix, MatrixBlock result) Left multiply with this column group
abstract void	leftMultByMatrix(MatrixBlock matrix, MatrixBlock result, int rl, int ru) Left multiply with this column group.
void	readFields(DataInput in) Deserialize column group from data input.
abstract AColGroup	replace(double pattern, double replace) Make a copy of the column group values, and replace all values that match pattern with replacement value.
abstract AColGroup	rightMultByMatrix(MatrixBlock right) Right matrix multiplication with this column group.
abstract AColGroup	scalarOperation(ScalarOperator op) Perform the specified scalar operation directly on the compressed column group, without decompressing individual cells if possible.
void	shiftColIndices(int offset) Shift all column indexes contained by an offset.
AColGroup	sliceColumn(int col) Slice out a single column from the column group.
AColGroup	sliceColumns(int cl, int cu) Slice out the columns within the range of cl and cu to remove the dictionary values related to these columns.
String	toString()
abstract void	tsmm(MatrixBlock ret, int nRows) Do a transposed self matrix multiplication on the left side t(x) %*% x.
abstract void	tsmmAColGroup(AColGroup other, MatrixBlock result) Matrix multiply with this other column group, but: 1.
abstract void	unaryAggregateOperations(AggregateUnaryOperator op, double[] c, int nRows, int rl, int ru) Unary Aggregate operator, since aggregate operators require new object output, the output becomes an uncompressed matrix.
void	write(DataOutput out) Serializes column group to data output.

Methods inherited from class java.lang.Object

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

Method Detail

getColIndices

public final int[] getColIndices()

Obtain the offsets of the columns in the matrix block that make up the group

Returns:

offsets of the columns in the matrix block that make up the group

getNumCols

public final int getNumCols()

Obtain the number of columns in this column group.

Returns:

number of columns in this column group

shiftColIndices

public final void shiftColIndices(int offset)

Shift all column indexes contained by an offset. This is used for rbind to combine compressed matrices. Since column indexes are reused between operations, we allocate a new list here to be safe

Parameters:

offset - The offset to move all columns

estimateInMemorySize

public long estimateInMemorySize()

Get the upper bound estimate of in memory allocation for the column group.

Returns:

an upper bound on the number of bytes used to store this ColGroup in memory.

decompressToBlock

public final void decompressToBlock(MatrixBlock target,
                                    int rl,
                                    int ru)

Decompress the contents of the column group without modifying nonZeroCount.

Parameters:

target - a matrix block where the columns covered by this column group have not yet been filled in.

rl - row lower

ru - row upper

write

public void write(DataOutput out)
           throws IOException

Serializes column group to data output.

Parameters:

out - data output

Throws:

IOException - if IOException occurs

readFields

public void readFields(DataInput in)
                throws IOException

Deserialize column group from data input.

Parameters:

in - data input

Throws:

IOException - if IOException occurs

getExactSizeOnDisk

public long getExactSizeOnDisk()

Returns the exact serialized size of column group. This can be used for example for buffer preallocation.

Returns:

exact serialized size for column group

leftMultByMatrix

public final void leftMultByMatrix(MatrixBlock matrix,
                                   MatrixBlock result)

Left multiply with this column group

Parameters:

matrix - The matrix to multiply with on the left

result - The result to output the values into, always dense for the purpose of the column groups parallelizing

sliceColumns

public final AColGroup sliceColumns(int cl,
                                    int cu)

Slice out the columns within the range of cl and cu to remove the dictionary values related to these columns. If the ColGroup slicing from does not contain any columns within the range null is returned.

Parameters:

cl - The lower bound of the columns to select

cu - The upper bound of the columns to select (not inclusive).

Returns:

A cloned Column Group, with a copied pointer to the old column groups index structure, but reduced dictionary and _columnIndexes correctly aligned with the expected sliced compressed matrix.

sliceColumn

public final AColGroup sliceColumn(int col)

Slice out a single column from the column group.

Parameters:

col - The column to slice, the column could potentially not be inside the column group

Returns:

A new column group that is a single column, if the column requested is not in this column group null is returned.

colSum

public static double[] colSum(List<AColGroup> groups,
                              double[] res,
                              int nRows)

Compute the column sum of the given list of groups

Parameters:

groups - The Groups to sum

res - The result to put the values into

nRows - The number of rows in the groups

Returns:

The given res list, where the sum of the column groups is added

getNumValues

public abstract int getNumValues()

Obtain number of distinct tuples in contained sets of values associated with this column group. If the column group is uncompressed the number or rows is returned.

Returns:

the number of distinct sets of values associated with the bitmaps in this column group

getCompType

public abstract AColGroup.CompressionType getCompType()

Obtain the compression type.

Returns:

How the elements of the column group are compressed.

decompressToBlock

public abstract void decompressToBlock(MatrixBlock target,
                                       int rl,
                                       int ru,
                                       int offT)

Decompress the contents of the column group without counting non zeros

Parameters:

target - a matrix block where the columns covered by this column group have not yet been filled in.

rl - row lower

ru - row upper

offT - Offset into target to assign from, this allows us to decompress into smaller matrices.

get

public abstract double get(int r,
                           int c)

Get the value at a global row/column position. In general this performs poorly compared to a normal lookup in a dense or sparse matrix.

Parameters:

r - row

c - column

Returns:

value at the row/column position

getValues

public abstract double[] getValues()

Get all the values in the colGroup. Note that this is only the stored values not the way they are stored. Making the output a list of values used in that colGroup not the actual full column.

Returns:

a double list of values.

getValuesAsBlock

public abstract MatrixBlock getValuesAsBlock()

Returns the ColGroup as a MatrixBlock. Used as a fall back solution in case a operation is not supported. Use in connection to getIfCountsType to get if the values are repeated.

Returns:

Matrix Block of the contained Values. Possibly contained in groups.

rightMultByMatrix

public abstract AColGroup rightMultByMatrix(MatrixBlock right)

Right matrix multiplication with this column group.

Parameters:

right - The matrixBlock on the right of this matrix multiplication

Returns:

The new Column Group that is the result of the matrix multiplication.

tsmm

public abstract void tsmm(MatrixBlock ret,
                          int nRows)

Do a transposed self matrix multiplication on the left side t(x) %*% x. but only with this column group. This gives better performance since there is no need to iterate through all the rows of the matrix, but the execution can be limited to its number of distinct values. Note it only calculate the upper triangle

Parameters:

ret - The return matrix block [numColumns x numColumns]

nRows - The number of rows in the column group

leftMultByMatrix

public abstract void leftMultByMatrix(MatrixBlock matrix,
                                      MatrixBlock result,
                                      int rl,
                                      int ru)

Left multiply with this column group.

Parameters:

matrix - The matrix to multiply with on the left

result - The result to output the values into, always dense for the purpose of the column groups parallelizing

rl - The row to begin the multiplication from on the lhs matrix

ru - The row to end the multiplication at on the lhs matrix

leftMultByAColGroup

public abstract void leftMultByAColGroup(AColGroup lhs,
                                         MatrixBlock result)

Left side matrix multiplication with a column group that is transposed.

Parameters:

lhs - The left hand side Column group to multiply with, the left hand side should be considered transposed.

result - The result matrix to insert the result of the multiplication into

tsmmAColGroup

public abstract void tsmmAColGroup(AColGroup other,
                                   MatrixBlock result)

Matrix multiply with this other column group, but: 1. Only output upper triangle values. 2. Multiply both ways with "this" being on the left and on the right. The second step is achievable by treating the initial multiplied matrix, and adding its values to the correct locations in the output.

Parameters:

other - The other Column group to multiply with

result - The result matrix to put the results into

scalarOperation

public abstract AColGroup scalarOperation(ScalarOperator op)

Perform the specified scalar operation directly on the compressed column group, without decompressing individual cells if possible.

Parameters:

op - operation to perform

Returns:

version of this column group with the operation applied

binaryRowOp

public abstract AColGroup binaryRowOp(BinaryOperator op,
                                      double[] v,
                                      boolean sparseSafe,
                                      boolean left)

Perform a binary row operation.

Parameters:

op - The operation to execute

v - The vector of values to apply, should be same length as dictionary length.

sparseSafe - True if the operation return 0 on all instances of values in v -- op(v[?], 0)

left - Specifies if the operation is executed on the left or right side of the values contained

Returns:

A updated column group with the new values.

unaryAggregateOperations

public abstract void unaryAggregateOperations(AggregateUnaryOperator op,
                                              double[] c,
                                              int nRows,
                                              int rl,
                                              int ru)

Unary Aggregate operator, since aggregate operators require new object output, the output becomes an uncompressed matrix.

Parameters:

op - The operator used

c - The output matrix block

nRows - The total number of rows in the Column Group

rl - The Starting Row to do aggregation from

ru - The last Row to do aggregation to (not included)

countNonZerosPerRow

public abstract void countNonZerosPerRow(int[] rnnz,
                                         int rl,
                                         int ru)

Count the number of non-zeros per row

Parameters:

rnnz - non-zeros per row

rl - row lower bound, inclusive

ru - row upper bound, exclusive

isLossy

public abstract boolean isLossy()

Is Lossy

Returns:

returns if the ColGroup is compressed in a lossy manner.

isDense

public abstract boolean isDense()

Is dense, signals that the entire column group is allocated an processed. This is useful in Row wise min and max for instance, to avoid having to scan through each row to look for empty rows. an example where it is true is DDC, Const and Uncompressed. examples where false is OLE and RLE.

Returns:

returns if the colGroup is allocated in a dense fashion.

getMin

public abstract double getMin()

Short hand method for getting minimum value contained in this column group.

Returns:

The minimum value contained in this ColumnGroup

getMax

public abstract double getMax()

Short hand method for getting maximum value contained in this column group.

Returns:

The maximum value contained in this ColumnGroup

copy

public abstract AColGroup copy()

Get a copy of this column group. Depending on which column group is copied it is a deep or shallow copy. If the primitives for the underlying column groups is Immutable then only shallow copies is performed.

Returns:

Get a copy of this column group.

containsValue

public abstract boolean containsValue(double pattern)

Detect if the column group contains a specific value.

Parameters:

pattern - The value to look for.

Returns:

boolean saying true if the value is contained.

getNumberNonZeros

public abstract long getNumberNonZeros(int nRows)

Get the number of nonZeros contained in this column group.

Parameters:

nRows - The number of rows in the column group, this is used for groups that does not contain information about how many rows they have.

Returns:

The nnz.

replace

public abstract AColGroup replace(double pattern,
                                  double replace)

Make a copy of the column group values, and replace all values that match pattern with replacement value.

Parameters:

pattern - The value to look for

replace - The value to replace the other value with

Returns:

A new Column Group, reusing the index structure but with new values.

computeColSums

public abstract void computeColSums(double[] c,
                                    int nRows)

Compute the column sum

Parameters:

c - The array to add the column sum to.

nRows - The number of rows in the column group.

toString

public String toString()

Overrides:

toString in class Object