use of org.knime.base.node.preproc.filter.row.rowfilter.IncludeFromNowOn in project knime-core by knime.
the class RowFilterIterator method getNextMatch.
/*
* returns the next row that is supposed to be returned or null if it met
* the end of it before.
*/
private DataRow getNextMatch() {
while (true) {
try {
m_exec.checkCanceled();
} catch (CanceledExecutionException cee) {
throw new RuntimeCanceledExecutionException(cee);
}
// we must not cause any trouble.
if (!m_orig.hasNext()) {
return null;
}
m_exec.setProgress(m_rowNumber / (double) m_totalCountInOrig);
DataRow next = m_orig.next();
if (m_includeRest) {
m_rowNumber++;
return next;
} else {
// consult the filter whether to include this row
try {
if (m_filter.matches(next, m_rowNumber)) {
return next;
}
// else fall through and get the next row from the orig
// table.
} catch (EndOfTableException eote) {
// filter: there are now more matching rows. Reached our
// EOT.
m_nextRow = null;
return null;
} catch (IncludeFromNowOn ifno) {
// filter: include all rows from now on
m_includeRest = true;
return next;
} finally {
m_rowNumber++;
}
}
}
}
use of org.knime.base.node.preproc.filter.row.rowfilter.IncludeFromNowOn in project knime-core by knime.
the class PartitionNodeModel method execute.
/**
* {@inheritDoc}
*/
@Override
protected BufferedDataTable[] execute(final BufferedDataTable[] inData, final ExecutionContext exec) throws CanceledExecutionException, Exception {
BufferedDataTable in = inData[0];
BufferedDataTable[] outs = new BufferedDataTable[2];
IRowFilter filter = getSamplingRowFilter(in, exec);
BufferedDataContainer firstOutCont = exec.createDataContainer(in.getDataTableSpec());
BufferedDataContainer secondOutCont = exec.createDataContainer(in.getDataTableSpec());
// floating point op. below
final double rowCount = in.size();
// one of the flags will be set if one of the exceptions below
// is thrown.
boolean putRestInOut1 = false;
boolean putRestInOut2 = false;
try {
int count = 0;
for (DataRow row : in) {
boolean matches = putRestInOut1;
try {
// conditional check, will call "matches" only if necessary
matches |= (!putRestInOut2 && filter.matches(row, count));
} catch (IncludeFromNowOn icf) {
assert !putRestInOut2;
putRestInOut1 = true;
matches = true;
} catch (EndOfTableException ete) {
assert !putRestInOut1;
putRestInOut2 = true;
matches = false;
}
if (matches) {
firstOutCont.addRowToTable(row);
} else {
secondOutCont.addRowToTable(row);
}
exec.setProgress(count / rowCount, "Processed row " + count + " (\"" + row.getKey() + "\")");
exec.checkCanceled();
count++;
}
} finally {
firstOutCont.close();
secondOutCont.close();
}
outs[0] = firstOutCont.getTable();
outs[1] = secondOutCont.getTable();
if (filter instanceof StratifiedSamplingRowFilter) {
int classCount = ((StratifiedSamplingRowFilter) filter).getClassCount();
if (classCount > outs[0].size()) {
setWarningMessage("Class column contains more classes (" + classCount + ") than sampled rows (" + outs[0].size() + ")");
}
}
return outs;
}
use of org.knime.base.node.preproc.filter.row.rowfilter.IncludeFromNowOn in project knime-core by knime.
the class RowFilter2PortNodeModel method execute.
/*
* The main work is done here
*
* @param rows total number of rows. Can be -1 if not available.
*/
private void execute(final RowInput in, final RowOutput match, final RowOutput miss, final long rows, final ExecutionContext exec) throws InterruptedException, CanceledExecutionException {
try {
long rowIdx = -1;
boolean allMatch = false;
boolean allMiss = false;
DataRow row;
while ((row = in.poll()) != null) {
rowIdx++;
if (rows > 0) {
exec.setProgress(rowIdx / (double) rows, "Adding row " + rowIdx + " of " + rows);
} else {
exec.setProgress("Adding row " + rowIdx + ".");
}
exec.checkCanceled();
if (allMatch) {
match.push(row);
continue;
}
if (allMiss) {
miss.push(row);
continue;
}
try {
if (m_rowFilter.matches(row, rowIdx)) {
match.push(row);
} else {
miss.push(row);
}
} catch (EndOfTableException eote) {
miss.push(row);
allMiss = true;
} catch (IncludeFromNowOn ifnoe) {
match.push(row);
allMatch = true;
}
}
} finally {
match.close();
miss.close();
}
}
use of org.knime.base.node.preproc.filter.row.rowfilter.IncludeFromNowOn in project knime-core by knime.
the class RowFilterNodeModel method execute.
private void execute(final RowInput inData, final RowOutput output, final ExecutionContext exec) throws Exception {
// in case the node was configured and the workflow is closed
// (and saved), the row filter isn't configured upon reloading.
// here, we give it a chance to configure itself (e.g. find the column
// index)
m_rowFilter.configure(inData.getDataTableSpec());
exec.setMessage("Searching first matching row...");
DataRow row;
int index = 0;
boolean isAlwaysExcluded = false;
boolean isAlwaysIncluded = false;
while ((row = inData.poll()) != null) {
boolean matches;
if (isAlwaysExcluded) {
matches = false;
} else if (isAlwaysIncluded) {
matches = true;
} else {
try {
matches = m_rowFilter.matches(row, index++);
} catch (EndOfTableException eot) {
break;
} catch (IncludeFromNowOn ifn) {
isAlwaysIncluded = true;
matches = true;
}
}
exec.checkCanceled();
if (matches) {
final int indexFinal = index;
final DataRow rowFinal = row;
exec.setMessage(() -> "Added row " + indexFinal + " (\"" + rowFinal.getKey() + "\")");
output.push(row);
}
}
inData.close();
output.close();
}
Aggregations