Examples with IOperatorDescriptor org.apache.hyracks.api.dataflow.IOperatorDescriptor used on opensource projects

Example 66 with IOperatorDescriptor

use of org.apache.hyracks.api.dataflow.IOperatorDescriptor in project asterixdb by apache.

the class CountOfCountsTest method countOfCountsSingleNC.

@Test
public void countOfCountsSingleNC() throws Exception {
    JobSpecification spec = new JobSpecification();
    FileSplit[] splits = new FileSplit[] { new ManagedFileSplit(NC2_ID, "data" + File.separator + "words.txt") };
    IFileSplitProvider splitProvider = new ConstantFileSplitProvider(splits);
    RecordDescriptor desc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer() });
    FileScanOperatorDescriptor csvScanner = new FileScanOperatorDescriptor(spec, splitProvider, new DelimitedDataTupleParserFactory(new IValueParserFactory[] { UTF8StringParserFactory.INSTANCE }, ','), desc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, csvScanner, NC2_ID);
    InMemorySortOperatorDescriptor sorter = new InMemorySortOperatorDescriptor(spec, new int[] { 0 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, desc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, sorter, NC2_ID);
    RecordDescriptor desc2 = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE });
    PreclusteredGroupOperatorDescriptor group = new PreclusteredGroupOperatorDescriptor(spec, new int[] { 0 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new CountFieldAggregatorFactory(true) }), desc2);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, group, NC2_ID);
    InMemorySortOperatorDescriptor sorter2 = new InMemorySortOperatorDescriptor(spec, new int[] { 1 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY) }, desc2);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, sorter2, NC2_ID);
    RecordDescriptor desc3 = new RecordDescriptor(new ISerializerDeserializer[] { IntegerSerializerDeserializer.INSTANCE, IntegerSerializerDeserializer.INSTANCE });
    PreclusteredGroupOperatorDescriptor group2 = new PreclusteredGroupOperatorDescriptor(spec, new int[] { 1 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY) }, new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new CountFieldAggregatorFactory(true) }), desc3);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, group2, NC2_ID);
    ResultSetId rsId = new ResultSetId(1);
    IOperatorDescriptor printer = new ResultWriterOperatorDescriptor(spec, rsId, true, false, ResultSerializerFactoryProvider.INSTANCE.getResultSerializerFactoryProvider());
    spec.addResultSetId(rsId);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC2_ID);
    IConnectorDescriptor conn1 = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(new int[] { 0 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
    spec.connect(conn1, csvScanner, 0, sorter, 0);
    IConnectorDescriptor conn2 = new OneToOneConnectorDescriptor(spec);
    spec.connect(conn2, sorter, 0, group, 0);
    IConnectorDescriptor conn3 = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(new int[] { 1 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
    spec.connect(conn3, group, 0, sorter2, 0);
    IConnectorDescriptor conn4 = new OneToOneConnectorDescriptor(spec);
    spec.connect(conn4, sorter2, 0, group2, 0);
    IConnectorDescriptor conn5 = new MToNBroadcastConnectorDescriptor(spec);
    spec.connect(conn5, group2, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
}

Example 67 with IOperatorDescriptor

use of org.apache.hyracks.api.dataflow.IOperatorDescriptor in project asterixdb by apache.

the class RTreeSecondaryIndexInsertOperatorTest method searchUpdatedSecondaryIndexTest.

@Test
public void searchUpdatedSecondaryIndexTest() throws Exception {
    JobSpecification spec = new JobSpecification();
    // build tuple
    ArrayTupleBuilder tb = new ArrayTupleBuilder(secondaryKeyFieldCount);
    DataOutput dos = tb.getDataOutput();
    tb.reset();
    DoubleSerializerDeserializer.INSTANCE.serialize(61.2894, dos);
    tb.addFieldEndOffset();
    DoubleSerializerDeserializer.INSTANCE.serialize(-149.624, dos);
    tb.addFieldEndOffset();
    DoubleSerializerDeserializer.INSTANCE.serialize(61.8894, dos);
    tb.addFieldEndOffset();
    DoubleSerializerDeserializer.INSTANCE.serialize(-149.024, dos);
    tb.addFieldEndOffset();
    ISerializerDeserializer[] keyRecDescSers = { DoubleSerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE };
    RecordDescriptor keyRecDesc = new RecordDescriptor(keyRecDescSers);
    ConstantTupleSourceOperatorDescriptor keyProviderOp = new ConstantTupleSourceOperatorDescriptor(spec, keyRecDesc, tb.getFieldEndOffsets(), tb.getByteArray(), tb.getSize());
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, keyProviderOp, NC1_ID);
    int[] keyFields = { 0, 1, 2, 3 };
    RTreeSearchOperatorDescriptor secondarySearchOp = new RTreeSearchOperatorDescriptor(spec, secondaryRecDesc, keyFields, true, true, secondaryHelperFactory, false, false, null, NoOpOperationCallbackFactory.INSTANCE, null, null, false);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, secondarySearchOp, NC1_ID);
    // fifth field from the tuples coming from secondary index
    int[] primaryLowKeyFields = { 4 };
    // fifth field from the tuples coming from secondary index
    int[] primaryHighKeyFields = { 4 };
    // search primary index
    BTreeSearchOperatorDescriptor primarySearchOp = new BTreeSearchOperatorDescriptor(spec, primaryRecDesc, primaryLowKeyFields, primaryHighKeyFields, true, true, primaryHelperFactory, false, false, null, NoOpOperationCallbackFactory.INSTANCE, null, null, false);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, primarySearchOp, NC1_ID);
    IFileSplitProvider outSplits = new ConstantFileSplitProvider(new FileSplit[] { createFile(nc1) });
    IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outSplits, ",");
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
    spec.connect(new OneToOneConnectorDescriptor(spec), keyProviderOp, 0, secondarySearchOp, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), secondarySearchOp, 0, primarySearchOp, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), primarySearchOp, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
}

Example 68 with IOperatorDescriptor

use of org.apache.hyracks.api.dataflow.IOperatorDescriptor in project asterixdb by apache.

the class RTreeSecondaryIndexScanOperatorTest method scanPrimaryIndexTest.

@Test
public void scanPrimaryIndexTest() throws Exception {
    JobSpecification spec = new JobSpecification();
    // build dummy tuple
    ArrayTupleBuilder tb = new ArrayTupleBuilder(secondaryKeyFieldCount);
    DataOutput dos = tb.getDataOutput();
    tb.reset();
    DoubleSerializerDeserializer.INSTANCE.serialize(0.0, dos);
    tb.addFieldEndOffset();
    DoubleSerializerDeserializer.INSTANCE.serialize(0.0, dos);
    tb.addFieldEndOffset();
    DoubleSerializerDeserializer.INSTANCE.serialize(0.0, dos);
    tb.addFieldEndOffset();
    DoubleSerializerDeserializer.INSTANCE.serialize(0.0, dos);
    tb.addFieldEndOffset();
    ISerializerDeserializer[] keyRecDescSers = { DoubleSerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE };
    RecordDescriptor keyRecDesc = new RecordDescriptor(keyRecDescSers);
    ConstantTupleSourceOperatorDescriptor keyProviderOp = new ConstantTupleSourceOperatorDescriptor(spec, keyRecDesc, tb.getFieldEndOffsets(), tb.getByteArray(), tb.getSize());
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, keyProviderOp, NC1_ID);
    int[] keyFields = null;
    RTreeSearchOperatorDescriptor secondarySearchOp = new RTreeSearchOperatorDescriptor(spec, secondaryRecDesc, keyFields, true, true, secondaryHelperFactory, false, false, null, NoOpOperationCallbackFactory.INSTANCE, null, null, false);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, secondarySearchOp, NC1_ID);
    IFileSplitProvider outSplits = new ConstantFileSplitProvider(new FileSplit[] { createFile(nc1) });
    IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outSplits, ",");
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
    spec.connect(new OneToOneConnectorDescriptor(spec), keyProviderOp, 0, secondarySearchOp, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), secondarySearchOp, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
}

Example 69 with IOperatorDescriptor

use of org.apache.hyracks.api.dataflow.IOperatorDescriptor in project asterixdb by apache.

the class RTreeSecondaryIndexStatsOperatorTest method showPrimaryIndexStats.

@Test
public void showPrimaryIndexStats() throws Exception {
    JobSpecification spec = new JobSpecification();
    TreeIndexStatsOperatorDescriptor secondaryStatsOp = new TreeIndexStatsOperatorDescriptor(spec, storageManager, secondaryHelperFactory);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, secondaryStatsOp, NC1_ID);
    IFileSplitProvider outSplits = new ConstantFileSplitProvider(new FileSplit[] { createFile(nc1) });
    IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outSplits, ",");
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
    spec.connect(new OneToOneConnectorDescriptor(spec), secondaryStatsOp, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
}

Example 70 with IOperatorDescriptor

use of org.apache.hyracks.api.dataflow.IOperatorDescriptor in project asterixdb by apache.

the class SecondaryInvertedIndexOperationsHelper method buildLoadingJobSpec.

@Override
public JobSpecification buildLoadingJobSpec() throws AlgebricksException {
    JobSpecification spec = RuntimeUtils.createJobSpecification(metadataProvider.getApplicationContext());
    JobId jobId = IndexUtil.bindJobEventListener(spec, metadataProvider);
    // Create dummy key provider for feeding the primary index scan.
    IOperatorDescriptor keyProviderOp = DatasetUtil.createDummyKeyProviderOp(spec, dataset, metadataProvider);
    // Create primary index scan op.
    IOperatorDescriptor primaryScanOp = DatasetUtil.createPrimaryIndexScanOp(spec, metadataProvider, dataset, jobId);
    IOperatorDescriptor sourceOp = primaryScanOp;
    boolean isEnforcingKeyTypes = index.isEnforcingKeyFileds();
    int numSecondaryKeys = index.getKeyFieldNames().size();
    if (isEnforcingKeyTypes && !enforcedItemType.equals(itemType)) {
        sourceOp = createCastOp(spec, dataset.getDatasetType());
        spec.connect(new OneToOneConnectorDescriptor(spec), primaryScanOp, 0, sourceOp, 0);
    }
    AlgebricksMetaOperatorDescriptor asterixAssignOp = createAssignOp(spec, numSecondaryKeys, secondaryRecDesc);
    // If any of the secondary fields are nullable, then add a select op
    // that filters nulls.
    AlgebricksMetaOperatorDescriptor selectOp = null;
    if (anySecondaryKeyIsNullable || isEnforcingKeyTypes) {
        selectOp = createFilterNullsSelectOp(spec, numSecondaryKeys, secondaryRecDesc);
    }
    // Create a tokenizer op.
    AbstractOperatorDescriptor tokenizerOp = createTokenizerOp(spec);
    // Sort by token + primary keys.
    ExternalSortOperatorDescriptor sortOp = createSortOp(spec, tokenKeyPairComparatorFactories, tokenKeyPairRecDesc);
    // Create secondary inverted index bulk load op.
    AbstractSingleActivityOperatorDescriptor invIndexBulkLoadOp = createInvertedIndexBulkLoadOp(spec);
    AlgebricksPartitionConstraintHelper.setPartitionConstraintInJobSpec(spec, invIndexBulkLoadOp, secondaryPartitionConstraint);
    AlgebricksMetaOperatorDescriptor metaOp = new AlgebricksMetaOperatorDescriptor(spec, 1, 0, new IPushRuntimeFactory[] { new SinkRuntimeFactory() }, new RecordDescriptor[] {});
    // Connect the operators.
    spec.connect(new OneToOneConnectorDescriptor(spec), keyProviderOp, 0, primaryScanOp, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), sourceOp, 0, asterixAssignOp, 0);
    if (anySecondaryKeyIsNullable || isEnforcingKeyTypes) {
        spec.connect(new OneToOneConnectorDescriptor(spec), asterixAssignOp, 0, selectOp, 0);
        spec.connect(new OneToOneConnectorDescriptor(spec), selectOp, 0, tokenizerOp, 0);
    } else {
        spec.connect(new OneToOneConnectorDescriptor(spec), asterixAssignOp, 0, tokenizerOp, 0);
    }
    spec.connect(new OneToOneConnectorDescriptor(spec), tokenizerOp, 0, sortOp, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), sortOp, 0, invIndexBulkLoadOp, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), invIndexBulkLoadOp, 0, metaOp, 0);
    spec.addRoot(metaOp);
    spec.setConnectorPolicyAssignmentPolicy(new ConnectorPolicyAssignmentPolicy());
    return spec;
}