Numeric table that provides methods to access data that is stored in the Compressed Sparse Row(CSR) data layout.
More...
◆ CSRNumericTable() [1/9]
Constructs sparse CSR numeric table from the array of doubles
- Parameters
-
context | Context to manage created CSR numeric table |
data | Array of values in the CSR layout |
colIndices | Array of column indices in the CSR layout. The values of indices are determined by the index base |
rowOffsets | Array of row indices in the CSR layout. The size of the array is nVectors+1. The first element is 0/1 in zero-/one-based indexing. The last element is ptr_size+0/1 in zero-/one-based indexing |
nFeatures | Number of columns in the corresponding dense table |
nVectors | Number of rows in the corresponding dense table |
indexing | Indexing scheme used to access data in the CSR layout Note: Present version of Intel(R) Data Analytics Acceleration Library supports 1-based indexing only |
◆ CSRNumericTable() [2/9]
Constructs sparse CSR numeric table from the array of doubles with 1-based indexing
- Parameters
-
context | Context to manage created CSR numeric table |
data | Array of values in the CSR layout |
colIndices | Array of column indices in the CSR layout. The values of indices are determined by the index base |
rowOffsets | Array of row indices in the CSR layout. The size of the array is nVectors+1. The first element is 0/1 in zero-/one-based indexing. The last element is ptr_size+0/1 in zero-/one-based indexing |
nFeatures | Number of columns in the corresponding dense table |
nVectors | Number of rows in the corresponding dense table |
◆ CSRNumericTable() [3/9]
Constructs sparse CSR numeric table from array of floats
- Parameters
-
context | Context to manage created CSR numeric table |
data | Array of values in the CSR layout |
colIndices | Array of column indices in the CSR layout. The values of indices are determined by the index base |
rowOffsets | Array of row indices in the CSR layout. The size of the array is nVectors+1. The first element is 0/1 in zero-/one-based indexing. The last element is ptr_size+0/1 in zero-/one-based indexing |
nFeatures | Number of columns in the corresponding dense table |
nVectors | Number of rows in the corresponding dense table |
indexing | Indexing scheme used to access data in the CSR layout Note: Present version of Intel(R) Data Analytics Acceleration Library supports 1-based indexing only |
◆ CSRNumericTable() [4/9]
Constructs sparse CSR numeric table from the array of float with 1-based indexing
- Parameters
-
context | Context to manage created CSR numeric table |
data | Array of values in the CSR layout |
colIndices | Array of column indices in the CSR layout. The values of indices are determined by the index base |
rowOffsets | Array of row indices in the CSR layout. The size of the array is nVectors+1. The first element is 0/1 in zero-/one-based indexing. The last element is ptr_size+0/1 in zero-/one-based indexing |
nFeatures | Number of columns in the corresponding dense table |
nVectors | Number of rows in the corresponding dense table |
◆ CSRNumericTable() [5/9]
Constructs sparse CSR numeric table from the array of integers
- Parameters
-
context | Context to manage created CSR numeric table |
data | Array of values in the CSR layout |
colIndices | Array of column indices in the CSR layout. The values of indices are determined by the index base |
rowOffsets | Array of row indices in the CSR layout. The size of the array is nVectors+1. The first element is 0/1 in zero-/one-based indexing. The last element is ptr_size+0/1 in zero-/one-based indexing |
nFeatures | Number of columns in the corresponding dense table |
nVectors | Number of rows in the corresponding dense table |
indexing | Indexing scheme used to access data in the CSR layout Note: Present version of Intel(R) Data Analytics Acceleration Library supports 1-based indexing only |
◆ CSRNumericTable() [6/9]
Constructs sparse CSR numeric table from the array of integers with 1-based indexing
- Parameters
-
context | Context to manage created CSR numeric table |
data | Array of values in the CSR layout |
colIndices | Array of column indices in the CSR layout. The values of indices are determined by the index base |
rowOffsets | Array of row indices in the CSR layout. The size of the array is nVectors+1. The first element is 0/1 in zero-/one-based indexing. The last element is ptr_size+0/1 in zero-/one-based indexing |
nFeatures | Number of columns in the corresponding dense table |
nVectors | Number of rows in the corresponding dense table |
◆ CSRNumericTable() [7/9]
Constructs sparse CSR numeric table from the array of longs
- Parameters
-
context | Context to manage created CSR numeric table |
data | Array of values in the CSR layout |
colIndices | Array of column indices in the CSR layout. The values of indices are determined by the index base |
rowOffsets | Array of row indices in the CSR layout. The size of the array is nVectors+1. The first element is 0/1 in zero-/one-based indexing. The last element is ptr_size+0/1 in zero-/one-based indexing |
nFeatures | Number of columns in the corresponding dense table |
nVectors | Number of rows in the corresponding dense table |
indexing | Indexing scheme used to access data in the CSR layout Note: Present version of Intel(R) Data Analytics Acceleration Library supports 1-based indexing only |
◆ CSRNumericTable() [8/9]
Constructs sparse CSR numeric table from the array of longs with 1-based indexing
- Parameters
-
context | Context to manage created CSR numeric table |
data | Array of values in the CSR layout |
colIndices | Array of column indices in the CSR layout. The values of indices are determined by the index base |
rowOffsets | Array of row indices in the CSR layout. The size of the array is nVectors+1. The first element is 0/1 in zero-/one-based indexing. The last element is ptr_size+0/1 in zero-/one-based indexing |
nFeatures | Number of columns in the corresponding dense table |
nVectors | Number of rows in the corresponding dense table |
◆ CSRNumericTable() [9/9]
Constructs homogeneous numeric table from C++ homogeneous numeric table
- Parameters
-
context | Context to manage the ALS algorithm |
cTable | Pointer to C++ numeric table |
◆ getBlockOfColumnValues() [1/3]
DoubleBuffer getBlockOfColumnValues |
( |
long |
featureIndex, |
|
|
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
DoubleBuffer |
buf |
|
) |
| |
Gets block of values for a given feature and returns it to java.nio.DoubleBuffer. This function needs to be defined by user in the subclass of this class.
- Parameters
-
featureIndex | Index of the feature |
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of values in the block |
buf | Buffer to store results |
- Returns
- Block of values of the feature packed into the DoubleBuffer
◆ getBlockOfColumnValues() [2/3]
FloatBuffer getBlockOfColumnValues |
( |
long |
featureIndex, |
|
|
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
FloatBuffer |
buf |
|
) |
| |
Gets block of values for a given feature and returns it to java.nio.FloatBuffer. This function needs to be defined by user in the subclass of this class.
- Parameters
-
featureIndex | Index of the feature |
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of values in the block |
buf | Buffer to store results |
- Returns
- Block of values of the feature packed into the FloatBuffer
◆ getBlockOfColumnValues() [3/3]
IntBuffer getBlockOfColumnValues |
( |
long |
featureIndex, |
|
|
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
IntBuffer |
buf |
|
) |
| |
Gets block of values for a given feature and returns it to java.nio.IntBuffer. This function needs to be defined by user in the subclass of this class.
- Parameters
-
featureIndex | Index of the feature |
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of values in the block |
buf | Buffer to store results |
- Returns
- Block of values of the feature packed into the IntBuffer
◆ getColIndicesArray()
long [] getColIndicesArray |
( |
| ) |
|
Gets data as an array of longs
- Returns
- Table data as an array of longs
◆ getDoubleArray()
double [] getDoubleArray |
( |
| ) |
|
Gets data as an array of doubles
- Returns
- Table data as an array of double
◆ getFloatArray()
float [] getFloatArray |
( |
| ) |
|
Gets data as an array of floats
- Returns
- Table data as an array of floats
◆ getLongArray()
Gets data as an array of longs
- Returns
- Table data as an array of longs
◆ getRowOffsetsArray()
long [] getRowOffsetsArray |
( |
| ) |
|
Gets data as an array of longs
- Returns
- Table data as an array of longs
◆ getSparseBlock()
long getSparseBlock |
( |
DataFeatureUtils.InternalNumType |
numType, |
|
|
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
ByteBuffer |
buf, |
|
|
LongBuffer |
colIndicesBuf, |
|
|
LongBuffer |
rowOffsetsBuf |
|
) |
| |
|
protected |
Reads block of rows from the table and returns it to java.nio.DoubleBuffer
- Parameters
-
numType | Type of data requested |
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of rows in the block |
buf | Buffer to store non-zero values |
colIndicesBuf | Buffer to store indices of the columns containing values from buf |
rowOffsetsBuf | Buffer to store row offsets of the values from buf |
- Returns
- Number of rows obtained from the table
◆ releaseBlockOfColumnValues() [1/3]
void releaseBlockOfColumnValues |
( |
long |
featureIndex, |
|
|
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
DoubleBuffer |
buf |
|
) |
| |
Transfers the values of a given feature from the input DoubleBuffer into a block of values of the feature in the table. This function needs to be defined by user in the subclass of this class.
- Parameters
-
featureIndex | Index of the feature |
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of values in the block |
buf | Input DoubleBuffer of size vectorNum |
◆ releaseBlockOfColumnValues() [2/3]
void releaseBlockOfColumnValues |
( |
long |
featureIndex, |
|
|
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
FloatBuffer |
buf |
|
) |
| |
Transfers the values of a given feature from the input FloatBuffer into a block of values of the feature in the table. This function needs to be defined by user in the subclass of this class.
- Parameters
-
featureIndex | Index of the feature |
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of values in the block |
buf | Input FloatBuffer of size vectorNum |
◆ releaseBlockOfColumnValues() [3/3]
void releaseBlockOfColumnValues |
( |
long |
featureIndex, |
|
|
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
IntBuffer |
buf |
|
) |
| |
Transfers the values of a given feature from the input IntBuffer into a block of values of the feature in the table. This function needs to be defined by user in the subclass of this class.
- Parameters
-
featureIndex | Index of the feature |
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of values in the block |
buf | Input IntBuffer of size vectorNum |
◆ releaseBlockOfRows() [1/3]
void releaseBlockOfRows |
( |
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
DoubleBuffer |
buf |
|
) |
| |
Transfers the data from the input DoubleBuffer into a block of table rows. This function needs to be defined by user in the subclass of this class.
- Parameters
-
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of rows in the block |
buf | Input DoubleBuffer with the capacity vectorNum * nColumns, where nColumns is the number of columns in the table |
◆ releaseBlockOfRows() [2/3]
void releaseBlockOfRows |
( |
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
FloatBuffer |
buf |
|
) |
| |
Transfers the data from the input FloatBuffer into a block of table rows. This function needs to be defined by user in the subclass of this class.
- Parameters
-
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of rows in the block |
buf | Input FloatBuffer with the capacity vectorNum * nColumns, where nColumns is the number of columns in the table |
◆ releaseBlockOfRows() [3/3]
void releaseBlockOfRows |
( |
long |
vectorIndex, |
|
|
long |
vectorNum, |
|
|
IntBuffer |
buf |
|
) |
| |
Transfers the data from the input IntBuffer into a block of table rows. This function needs to be defined by user in the subclass of this class.
- Parameters
-
vectorIndex | Index of the first row to include into the block |
vectorNum | Number of rows in the block |
buf | Input IntBuffer with the capacity vectorNum * nColumns, where nColumns is the number of columns in the table |
|
class | Indexing |
| Indexing scheme used for accessing the data in CSR layout More...
|
|
The documentation for this class was generated from the following file: