| package edu.uci.ics.asterix.runtime.evaluators.functions; |
| |
| import java.io.ByteArrayInputStream; |
| import java.io.DataInputStream; |
| import java.io.DataOutput; |
| import java.util.regex.Matcher; |
| import java.util.regex.Pattern; |
| |
| import edu.uci.ics.asterix.common.utils.UTF8CharSequence; |
| import edu.uci.ics.asterix.formats.nontagged.AqlBinaryComparatorFactoryProvider; |
| import edu.uci.ics.asterix.formats.nontagged.AqlSerializerDeserializerProvider; |
| import edu.uci.ics.asterix.om.base.AString; |
| import edu.uci.ics.asterix.om.functions.AsterixBuiltinFunctions; |
| import edu.uci.ics.asterix.om.functions.IFunctionDescriptor; |
| import edu.uci.ics.asterix.om.functions.IFunctionDescriptorFactory; |
| import edu.uci.ics.asterix.om.types.BuiltinType; |
| import edu.uci.ics.asterix.runtime.evaluators.base.AbstractScalarFunctionDynamicDescriptor; |
| import edu.uci.ics.hyracks.algebricks.common.exceptions.AlgebricksException; |
| import edu.uci.ics.hyracks.algebricks.core.algebra.functions.FunctionIdentifier; |
| import edu.uci.ics.hyracks.algebricks.runtime.base.ICopyEvaluator; |
| import edu.uci.ics.hyracks.algebricks.runtime.base.ICopyEvaluatorFactory; |
| import edu.uci.ics.hyracks.api.dataflow.value.IBinaryComparator; |
| import edu.uci.ics.hyracks.api.dataflow.value.ISerializerDeserializer; |
| import edu.uci.ics.hyracks.api.exceptions.HyracksDataException; |
| import edu.uci.ics.hyracks.data.std.api.IDataOutputProvider; |
| import edu.uci.ics.hyracks.data.std.util.ArrayBackedValueStorage; |
| import edu.uci.ics.hyracks.data.std.util.ByteArrayAccessibleOutputStream; |
| |
| /** |
| * @author Xiaoyu Ma |
| */ |
| public class StringMatchesDescriptor extends AbstractScalarFunctionDynamicDescriptor { |
| private static final long serialVersionUID = 1L; |
| |
| public static final IFunctionDescriptorFactory FACTORY = new IFunctionDescriptorFactory() { |
| public IFunctionDescriptor createFunctionDescriptor() { |
| return new StringMatchesDescriptor(); |
| } |
| }; |
| |
| @Override |
| public ICopyEvaluatorFactory createEvaluatorFactory(final ICopyEvaluatorFactory[] args) throws AlgebricksException { |
| |
| return new ICopyEvaluatorFactory() { |
| private static final long serialVersionUID = 1L; |
| |
| @Override |
| public ICopyEvaluator createEvaluator(IDataOutputProvider output) throws AlgebricksException { |
| |
| DataOutput dout = output.getDataOutput(); |
| |
| return new AbstractBinaryStringBoolEval(dout, args[0], args[1]) { |
| |
| private Pattern pattern = null; |
| private Matcher matcher = null; |
| private ByteArrayAccessibleOutputStream lastPattern = new ByteArrayAccessibleOutputStream(); |
| private IBinaryComparator strComp = AqlBinaryComparatorFactoryProvider.INSTANCE |
| .getBinaryComparatorFactory(BuiltinType.ASTRING, true).createBinaryComparator(); |
| private UTF8CharSequence carSeq = new UTF8CharSequence(); |
| |
| @SuppressWarnings("unchecked") |
| private ISerializerDeserializer<AString> stringSerde = AqlSerializerDeserializerProvider.INSTANCE |
| .getSerializerDeserializer(BuiltinType.ASTRING); |
| |
| @Override |
| protected boolean compute(byte[] lBytes, int lLen, int lStart, byte[] rBytes, int rLen, int rStart, |
| ArrayBackedValueStorage array0, ArrayBackedValueStorage array1) throws AlgebricksException { |
| try { |
| boolean newPattern = false; |
| if (pattern == null) { |
| newPattern = true; |
| } else { |
| int c = strComp.compare(rBytes, rStart, rLen, lastPattern.getByteArray(), 0, |
| lastPattern.size()); |
| if (c != 0) { |
| newPattern = true; |
| } |
| } |
| if (newPattern) { |
| lastPattern.reset(); |
| lastPattern.write(rBytes, rStart, rLen); |
| // ! object creation ! |
| DataInputStream di = new DataInputStream(new ByteArrayInputStream( |
| lastPattern.getByteArray())); |
| AString strPattern = (AString) stringSerde.deserialize(di); |
| // pattern = Pattern.compile(toRegex(strPattern)); |
| pattern = Pattern.compile(strPattern.getStringValue()); |
| } |
| |
| carSeq.reset(array0, 1); |
| if (newPattern) { |
| matcher = pattern.matcher(carSeq); |
| } else { |
| matcher.reset(carSeq); |
| } |
| return matcher.find(); |
| } catch (HyracksDataException e) { |
| throw new AlgebricksException(e); |
| } |
| } |
| |
| }; |
| } |
| }; |
| } |
| |
| @Override |
| public FunctionIdentifier getIdentifier() { |
| return AsterixBuiltinFunctions.STRING_MATCHES; |
| } |
| } |