Examples with FileScanOperatorDescriptor org.apache.hyracks.dataflow.std.file.FileScanOperatorDescriptor used on opensource projects

Example 36 with FileScanOperatorDescriptor

use of org.apache.hyracks.dataflow.std.file.FileScanOperatorDescriptor in project asterixdb by apache.

the class Sort method createJob.

private static JobSpecification createJob(FileSplit[] ordersSplits, FileSplit[] outputSplit, String memBufferAlg, int frameLimit, int frameSize, int limit, boolean usingHeapSorter) {
    JobSpecification spec = new JobSpecification();
    spec.setFrameSize(frameSize);
    IFileSplitProvider ordersSplitProvider = new ConstantFileSplitProvider(ordersSplits);
    FileScanOperatorDescriptor ordScanner = new FileScanOperatorDescriptor(spec, ordersSplitProvider, new DelimitedDataTupleParserFactory(orderParserFactories, '|'), ordersDesc);
    createPartitionConstraint(spec, ordScanner, ordersSplits);
    AbstractSorterOperatorDescriptor sorter;
    if (usingHeapSorter && limit < Integer.MAX_VALUE) {
        sorter = new TopKSorterOperatorDescriptor(spec, frameLimit, limit, SortFields, null, SortFieldsComparatorFactories, ordersDesc);
    } else {
        if (memBufferAlg.equalsIgnoreCase("bestfit")) {
            sorter = new ExternalSortOperatorDescriptor(spec, frameLimit, SortFields, null, SortFieldsComparatorFactories, ordersDesc, Algorithm.MERGE_SORT, EnumFreeSlotPolicy.SMALLEST_FIT, limit);
        } else if (memBufferAlg.equalsIgnoreCase("biggestfit")) {
            sorter = new ExternalSortOperatorDescriptor(spec, frameLimit, SortFields, null, SortFieldsComparatorFactories, ordersDesc, Algorithm.MERGE_SORT, EnumFreeSlotPolicy.BIGGEST_FIT, limit);
        } else {
            sorter = new ExternalSortOperatorDescriptor(spec, frameLimit, SortFields, null, SortFieldsComparatorFactories, ordersDesc, Algorithm.MERGE_SORT, EnumFreeSlotPolicy.LAST_FIT, limit);
        }
    }
    createPartitionConstraint(spec, sorter, ordersSplits);
    IFileSplitProvider outputSplitProvider = new ConstantFileSplitProvider(outputSplit);
    IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outputSplitProvider, "|");
    createPartitionConstraint(spec, printer, outputSplit);
    spec.connect(new OneToOneConnectorDescriptor(spec), ordScanner, 0, sorter, 0);
    spec.connect(new MToNPartitioningMergingConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(SortFields, orderBinaryHashFunctionFactories), SortFields, SortFieldsComparatorFactories, new UTF8StringNormalizedKeyComputerFactory()), sorter, 0, printer, 0);
    spec.addRoot(printer);
    return spec;
}

Example 37 with FileScanOperatorDescriptor

use of org.apache.hyracks.dataflow.std.file.FileScanOperatorDescriptor in project asterixdb by apache.

the class LocalityAwareConnectorTest method localityAwareAggregationTest.

/**
     * Test of aggregations using locality aware connector. The output two files should be the
     * same, each of which is the aggregation of two copies of the lineitem.tbl.
     * Note that if the hashing connector is not working correctly, the two files may be different. This
     * also means that even the output files are the same, the hashing may have other problems.
     *
     * @throws Exception
     */
@Test
public void localityAwareAggregationTest() throws Exception {
    JobSpecification spec = new JobSpecification();
    FileScanOperatorDescriptor csvScanner = new FileScanOperatorDescriptor(spec, splitProvider, tupleParserFactory, desc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, csvScanner, "asterix-001", "asterix-002", "asterix-003", "asterix-004");
    RecordDescriptor outputRec = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE, IntegerSerializerDeserializer.INSTANCE, FloatSerializerDeserializer.INSTANCE });
    int[] keyFields = new int[] { 0 };
    int tableSize = 8;
    ExternalGroupOperatorDescriptor grouper = new ExternalGroupOperatorDescriptor(spec, tableSize, fileSize, keyFields, fileSize / spec.getFrameSize() + 1, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new UTF8StringNormalizedKeyComputerFactory(), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(1, false), new IntSumFieldAggregatorFactory(3, false), new FloatSumFieldAggregatorFactory(5, false) }), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(1, false), new IntSumFieldAggregatorFactory(2, false), new FloatSumFieldAggregatorFactory(3, false) }), outputRec, outputRec, new HashSpillableTableFactory(new IBinaryHashFunctionFamily[] { UTF8StringBinaryHashFunctionFamily.INSTANCE }));
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, grouper, "asterix-005", "asterix-006");
    BitSet nodemap = new BitSet(8);
    nodemap.set(0);
    nodemap.set(2);
    nodemap.set(5);
    nodemap.set(7);
    IConnectorDescriptor conn1 = new LocalityAwareMToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(keyFields, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }), new HashtableLocalityMap(nodemap));
    spec.connect(conn1, csvScanner, 0, grouper, 0);
    AbstractSingleActivityOperatorDescriptor printer = getPrinter(spec, "localityAwareSumInmemGroupTest");
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, "asterix-005", "asterix-006");
    IConnectorDescriptor conn2 = new OneToOneConnectorDescriptor(spec);
    spec.connect(conn2, grouper, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
}

Example 38 with FileScanOperatorDescriptor

use of org.apache.hyracks.dataflow.std.file.FileScanOperatorDescriptor in project asterixdb by apache.

the class LocalityAwareConnectorTest method globalPartitionAggregationTest.

/**
     * Test for locality aware connector, using the global hashing node mapper. This should have
     * the exactly the same result as using {@link MToNPartitioningConnectorDescriptor}.
     *
     * @throws Exception
     */
@Test
public void globalPartitionAggregationTest() throws Exception {
    JobSpecification spec = new JobSpecification();
    FileScanOperatorDescriptor csvScanner = new FileScanOperatorDescriptor(spec, splitProvider, tupleParserFactory, desc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, csvScanner, "asterix-001", "asterix-002", "asterix-003", "asterix-004");
    RecordDescriptor outputRec = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE, IntegerSerializerDeserializer.INSTANCE, FloatSerializerDeserializer.INSTANCE });
    int[] keyFields = new int[] { 0 };
    int tableSize = 8;
    ExternalGroupOperatorDescriptor grouper = new ExternalGroupOperatorDescriptor(spec, tableSize, fileSize, keyFields, fileSize / spec.getFrameSize() + 1, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new UTF8StringNormalizedKeyComputerFactory(), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(1, false), new IntSumFieldAggregatorFactory(3, false), new FloatSumFieldAggregatorFactory(5, false) }), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(1, false), new IntSumFieldAggregatorFactory(2, false), new FloatSumFieldAggregatorFactory(3, false) }), outputRec, outputRec, new HashSpillableTableFactory(new IBinaryHashFunctionFamily[] { UTF8StringBinaryHashFunctionFamily.INSTANCE }));
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, grouper, "asterix-005", "asterix-006");
    IConnectorDescriptor conn1 = new LocalityAwareMToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(keyFields, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }), new GlobalHashingLocalityMap());
    spec.connect(conn1, csvScanner, 0, grouper, 0);
    AbstractSingleActivityOperatorDescriptor printer = getPrinter(spec, "localityAwareSumInmemGroupTest");
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, "asterix-005", "asterix-006");
    IConnectorDescriptor conn2 = new OneToOneConnectorDescriptor(spec);
    spec.connect(conn2, grouper, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
}

Example 39 with FileScanOperatorDescriptor

use of org.apache.hyracks.dataflow.std.file.FileScanOperatorDescriptor 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 40 with FileScanOperatorDescriptor

use of org.apache.hyracks.dataflow.std.file.FileScanOperatorDescriptor in project asterixdb by apache.

the class AggregationTest method multiKeySumExtGroupTest.

@Test
public void multiKeySumExtGroupTest() throws Exception {
    JobSpecification spec = new JobSpecification();
    FileScanOperatorDescriptor csvScanner = new FileScanOperatorDescriptor(spec, splitProvider, tupleParserFactory, desc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, csvScanner, NC2_ID);
    RecordDescriptor outputRec = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE, IntegerSerializerDeserializer.INSTANCE });
    int[] keyFields = new int[] { 8, 0 };
    int frameLimits = 5;
    int tableSize = 8;
    long fileSize = frameLimits * spec.getFrameSize();
    ExternalGroupOperatorDescriptor grouper = new ExternalGroupOperatorDescriptor(spec, tableSize, fileSize, keyFields, frameLimits, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new UTF8StringNormalizedKeyComputerFactory(), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(1, false), new IntSumFieldAggregatorFactory(3, false) }), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(2, false), new IntSumFieldAggregatorFactory(3, false) }), outputRec, outputRec, new HashSpillableTableFactory(new IBinaryHashFunctionFamily[] { UTF8StringBinaryHashFunctionFamily.INSTANCE, UTF8StringBinaryHashFunctionFamily.INSTANCE }));
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, grouper, NC2_ID, NC1_ID);
    IConnectorDescriptor conn1 = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(keyFields, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY), PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
    spec.connect(conn1, csvScanner, 0, grouper, 0);
    AbstractSingleActivityOperatorDescriptor printer = getPrinter(spec, "multiKeySumExtGroupTest");
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC2_ID, NC1_ID);
    IConnectorDescriptor conn2 = new OneToOneConnectorDescriptor(spec);
    spec.connect(conn2, grouper, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
}