use of org.knime.base.data.aggregation.dialogutil.pattern.PatternAggregator in project knime-core by knime.
the class GroupByNodeModel method getAggregators.
/**
* Creates a {@link List} with all {@link ColumnAggregator}s to use based on the given input settings.
* Columns are only added once for the different aggregator types in the order they are added to the function
* e.g. all column that are handled by one of the given {@link ColumnAggregator} are ignored by the
* pattern and data type based aggregator all columns that are handled by one of the pattern based aggregators
* is ignored by the data type based aggregators.
* @param inputSpec the {@link DataTableSpec} of the input table
* @param groupColumns the columns to group by
* @param columnAggregators the manually added {@link ColumnAggregator}s
* @param patternAggregators the {@link PatternAggregator}s
* @param dataTypeAggregators the {@link DataTypeAggregator}s
* @param invalidColAggrs empty {@link List} that is filled with the invalid column aggregators can be
* <code>null</code>
* @return the list of all {@link ColumnAggregator}s to use based on the given aggregator
* @since 2.11
*/
public static List<ColumnAggregator> getAggregators(final DataTableSpec inputSpec, final Collection<String> groupColumns, final List<ColumnAggregator> columnAggregators, final Collection<PatternAggregator> patternAggregators, final Collection<DataTypeAggregator> dataTypeAggregators, final List<ColumnAggregator> invalidColAggrs) {
final List<ColumnAggregator> columnAggregators2Use = new ArrayList<>(columnAggregators.size());
final Set<String> usedColNames = new HashSet<>(inputSpec.getNumColumns());
usedColNames.addAll(groupColumns);
for (final ColumnAggregator colAggr : columnAggregators) {
final String originalColName = colAggr.getOriginalColName();
final DataColumnSpec colSpec = inputSpec.getColumnSpec(originalColName);
if (colSpec != null && colAggr.getOriginalDataType().isASuperTypeOf(colSpec.getType())) {
usedColNames.add(originalColName);
columnAggregators2Use.add(colAggr);
} else {
if (invalidColAggrs != null) {
invalidColAggrs.add(colAggr);
}
}
}
if (inputSpec.getNumColumns() > usedColNames.size() && !patternAggregators.isEmpty()) {
for (final DataColumnSpec spec : inputSpec) {
if (!usedColNames.contains(spec.getName())) {
for (final PatternAggregator patternAggr : patternAggregators) {
Pattern pattern = patternAggr.getRegexPattern();
if (pattern != null && pattern.matcher(spec.getName()).matches() && patternAggr.isCompatible(spec)) {
final ColumnAggregator colAggregator = new ColumnAggregator(spec, patternAggr.getMethodTemplate(), patternAggr.inclMissingCells());
columnAggregators2Use.add(colAggregator);
usedColNames.add(spec.getName());
}
}
}
}
}
// check if some columns are left
if (inputSpec.getNumColumns() > usedColNames.size() && !dataTypeAggregators.isEmpty()) {
for (final DataColumnSpec spec : inputSpec) {
if (!usedColNames.contains(spec.getName())) {
final DataType dataType = spec.getType();
for (final DataTypeAggregator typeAggregator : dataTypeAggregators) {
if (typeAggregator.isCompatibleType(dataType)) {
final ColumnAggregator colAggregator = new ColumnAggregator(spec, typeAggregator.getMethodTemplate(), typeAggregator.inclMissingCells());
columnAggregators2Use.add(colAggregator);
usedColNames.add(spec.getName());
}
}
}
}
}
return columnAggregators2Use;
}
use of org.knime.base.data.aggregation.dialogutil.pattern.PatternAggregator in project knime-core by knime.
the class GroupByNodeModel method validateSettings.
/**
* {@inheritDoc}
*/
@Override
protected void validateSettings(final NodeSettingsRO settings) throws InvalidSettingsException {
m_groupByCols.validateSettings(settings);
// FIX bug 5040: potential problem with clone settings method when in-/exclude list contain same elements
final SettingsModelFilterString tmpSett = new SettingsModelFilterString(CFG_GROUP_BY_COLUMNS);
tmpSett.loadSettingsFrom(settings);
final List<String> groupByCols = tmpSett.getIncludeList();
m_maxUniqueValues.validateSettings(settings);
m_enableHilite.validateSettings(settings);
// with Knime 2.0 as well as the naming policy
try {
final List<ColumnAggregator> aggregators = ColumnAggregator.loadColumnAggregators(settings);
final List<DataTypeAggregator> typeAggregators = new LinkedList<>();
final List<PatternAggregator> patternAggregators = new LinkedList<>();
try {
patternAggregators.addAll(PatternAggregator.loadAggregators(settings, CFG_PATTERN_AGGREGATORS));
typeAggregators.addAll(DataTypeAggregator.loadAggregators(settings, CFG_DATA_TYPE_AGGREGATORS));
} catch (InvalidSettingsException e) {
// introduced in 2.11
}
if (groupByCols.isEmpty() && aggregators.isEmpty() && patternAggregators.isEmpty() && typeAggregators.isEmpty()) {
throw new IllegalArgumentException("Please select at least one group column or aggregation option");
}
ColumnNamePolicy namePolicy;
try {
final String policyLabel = ((SettingsModelString) m_columnNamePolicy.createCloneWithValidatedValue(settings)).getStringValue();
namePolicy = ColumnNamePolicy.getPolicy4Label(policyLabel);
} catch (final InvalidSettingsException e) {
namePolicy = compGetColumnNamePolicy(settings);
}
checkDuplicateAggregators(namePolicy, aggregators);
} catch (final InvalidSettingsException e) {
// these settings are prior Knime 2.0 and can't contain
// a column several times
} catch (final IllegalArgumentException e) {
throw new InvalidSettingsException(e.getMessage());
}
}
use of org.knime.base.data.aggregation.dialogutil.pattern.PatternAggregator in project knime-core by knime.
the class MovingAggregationNodeModel method validateSettings.
/**
* {@inheritDoc}
*/
@Override
protected void validateSettings(final NodeSettingsRO settings) throws InvalidSettingsException {
m_winLength.validateSettings(settings);
m_windowType.validateSettings(settings);
m_cumulativeComputing.validateSettings(settings);
m_handleMissings.validateSettings(settings);
m_removeRetainedCols.validateSettings(settings);
m_maxUniqueVals.validateSettings(settings);
m_valueDelimiter.validateSettings(settings);
final List<ColumnAggregator> aggregators = ColumnAggregator.loadColumnAggregators(settings);
final List<DataTypeAggregator> typeAggregators = new LinkedList<>();
final List<PatternAggregator> regexAggregators = new LinkedList<>();
try {
regexAggregators.addAll(PatternAggregator.loadAggregators(settings, CFG_PATTERN_AGGREGATORS));
typeAggregators.addAll(DataTypeAggregator.loadAggregators(settings, CFG_DATA_TYPE_AGGREGATORS));
} catch (InvalidSettingsException e) {
// introduced in 2.11
}
if (aggregators.isEmpty() && regexAggregators.isEmpty() && typeAggregators.isEmpty()) {
throw new IllegalArgumentException("Please select at least one aggregation option");
}
final String policyLabel = ((SettingsModelString) m_columnNamePolicy.createCloneWithValidatedValue(settings)).getStringValue();
final ColumnNamePolicy namePolicy = ColumnNamePolicy.getPolicy4Label(policyLabel);
try {
GroupByNodeModel.checkDuplicateAggregators(namePolicy, aggregators);
} catch (IllegalArgumentException e) {
throw new InvalidSettingsException(e.getMessage());
}
final boolean removeAggrCols = ((SettingsModelBoolean) m_removeAggregationCols.createCloneWithValidatedValue(settings)).getBooleanValue();
if (ColumnNamePolicy.KEEP_ORIGINAL_NAME.equals(namePolicy) && !removeAggrCols) {
throw new InvalidSettingsException("'" + ColumnNamePolicy.KEEP_ORIGINAL_NAME.getLabel() + "' option only valid if aggregation columns are filtered");
}
}
Aggregations