EDOBE XDOM PMML Manual de usuario descargar pdf (Pagina 30)

* is executed.

public void execute(CNKProcJavaTransform proc) {

for (int i=0; i<getNumInputs(); i++) {

proc.copyData(i, 0, 0, i, 0, 0,

proc.getChunkInputRows(i),

getInputMetaData(i).getNumColumns());

}

This example file does not start with a package statement, so the

FirstCopyEngineNode class will be part of the unnamed default

package. We import the files in the

com.insightful.miner and

com.insightful.cnkjava packages. We could be more selective

regarding our imports and only import the classes that we actually

use.

The

FirstCopyEngineNode class will use a Java transform, so it is

defined to implement

CNKProcJavaTransformExec. In procCreate()

we create a

CNKProcJavaTransform object that will call the

FirstCopyEngineNode.execute() method once for each block of data.

calculateOutputMetaData() we copy the metadata for each input

to the corresponding output, since the names and types of the output

columns are the same as the corresponding input.

execute() we loop over the inputs and use the

CNKProcJavaTransform method copyData() to copy the specified

range of rows. The arguments for this function are:

public void copyData(int outputNum, int outputFirstCol,

int outputFirstRow, int inputNum, int inputFirstCol,

int inputFirstRow, int numRows, int numColumns)

For each input, we get the number of rows in the current chunk from

the

CNKProcJavaTransform and the number of columns from the

input’s

XTMetaData.

1 2 ... 25 26 27 28 29 30 31 32 33 34 35 ... 97 98

Sin comentarios

EDOBE XDOM PMML Manual de usuario Pagina 30