Examples with JobSpecification org.apache.hyracks.api.job.JobSpecification used on opensource projects

Example 76 with JobSpecification

use of org.apache.hyracks.api.job.JobSpecification in project asterixdb by apache.

the class Join method createJob.

private static JobSpecification createJob(FileSplit[] customerSplits, FileSplit[] orderSplits, FileSplit[] resultSplits, int numJoinPartitions, String algo, int graceInputSize, int graceRecordsPerFrame, double graceFactor, int memSize, int tableSize, boolean hasGroupBy, int frameSize) throws HyracksDataException {
    JobSpecification spec = new JobSpecification(frameSize);
    IFileSplitProvider custSplitsProvider = new ConstantFileSplitProvider(customerSplits);
    long custFileSize = 0;
    for (int i = 0; i < customerSplits.length; i++) {
        custFileSize += customerSplits[i].getFile(null).length();
    }
    IFileSplitProvider ordersSplitsProvider = new ConstantFileSplitProvider(orderSplits);
    long orderFileSize = 0;
    for (int i = 0; i < orderSplits.length; i++) {
        orderFileSize += orderSplits[i].getFile(null).length();
    }
    FileScanOperatorDescriptor ordScanner = new FileScanOperatorDescriptor(spec, ordersSplitsProvider, new DelimitedDataTupleParserFactory(orderParserFactories, '|'), Common.ordersDesc);
    createPartitionConstraint(spec, ordScanner, orderSplits);
    FileScanOperatorDescriptor custScanner = new FileScanOperatorDescriptor(spec, custSplitsProvider, new DelimitedDataTupleParserFactory(custParserFactories, '|'), Common.custDesc);
    createPartitionConstraint(spec, custScanner, customerSplits);
    IOperatorDescriptor join;
    if ("nestedloop".equalsIgnoreCase(algo)) {
        join = new NestedLoopJoinOperatorDescriptor(spec, new JoinComparatorFactory(PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), 0, 1), Common.custOrderJoinDesc, memSize, false, null);
    } else if ("inmem".equalsIgnoreCase(algo)) {
        join = new InMemoryHashJoinOperatorDescriptor(spec, new int[] { 0 }, new int[] { 1 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, Common.custOrderJoinDesc, tableSize, null, memSize * frameSize);
    } else if ("hybrid".equalsIgnoreCase(algo)) {
        join = new OptimizedHybridHashJoinOperatorDescriptor(spec, memSize, graceInputSize, graceFactor, new int[] { 0 }, new int[] { 1 }, new IBinaryHashFunctionFamily[] { UTF8StringBinaryHashFunctionFamily.INSTANCE }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, Common.custOrderJoinDesc, new JoinComparatorFactory(PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), 0, 1), new JoinComparatorFactory(PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), 1, 0), null);
    } else {
        System.err.println("unknown algorithm:" + algo);
        return null;
    }
    PartitionConstraintHelper.addPartitionCountConstraint(spec, join, numJoinPartitions);
    IConnectorDescriptor ordJoinConn = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(new int[] { 1 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
    spec.connect(ordJoinConn, ordScanner, 0, join, 1);
    IConnectorDescriptor custJoinConn = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(new int[] { 0 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
    spec.connect(custJoinConn, custScanner, 0, join, 0);
    IOperatorDescriptor endingOp = join;
    if (hasGroupBy) {
        RecordDescriptor groupResultDesc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE });
        ExternalGroupOperatorDescriptor gby = new ExternalGroupOperatorDescriptor(spec, tableSize, custFileSize + orderFileSize, new int[] { 6 }, memSize, 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) }), groupResultDesc, groupResultDesc, new HashSpillableTableFactory(new IBinaryHashFunctionFamily[] { UTF8StringBinaryHashFunctionFamily.INSTANCE }));
        createPartitionConstraint(spec, gby, resultSplits);
        IConnectorDescriptor joinGroupConn = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(new int[] { 6 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
        spec.connect(joinGroupConn, join, 0, gby, 0);
        endingOp = gby;
    }
    IFileSplitProvider outSplitProvider = new ConstantFileSplitProvider(resultSplits);
    //FrameFileWriterOperatorDescriptor writer = new FrameFileWriterOperatorDescriptor(spec, outSplitProvider);
    IOperatorDescriptor writer = new PlainFileWriterOperatorDescriptor(spec, outSplitProvider, "|");
    createPartitionConstraint(spec, writer, resultSplits);
    IConnectorDescriptor endingPrinterConn = new OneToOneConnectorDescriptor(spec);
    spec.connect(endingPrinterConn, endingOp, 0, writer, 0);
    spec.addRoot(writer);
    return spec;
}

Example 77 with JobSpecification

use of org.apache.hyracks.api.job.JobSpecification in project asterixdb by apache.

the class WordCountMain method main.

public static void main(String[] args) throws Exception {
    Options options = new Options();
    CmdLineParser parser = new CmdLineParser(options);
    parser.parseArgument(args);
    IHyracksClientConnection hcc = new HyracksConnection(options.host, options.port);
    JobSpecification job = createJob(parseFileSplits(options.inFileSplits), parseFileSplits(options.outFileSplits), options.algo, options.htSize, options.memFrameLimit, options.format, options.frameSize);
    long start = System.currentTimeMillis();
    JobId jobId = hcc.startJob(job, options.runtimeProfiling ? EnumSet.of(JobFlag.PROFILE_RUNTIME) : EnumSet.noneOf(JobFlag.class));
    hcc.waitForCompletion(jobId);
    long end = System.currentTimeMillis();
    System.err.println(start + " " + end + " " + (end - start));
}

Example 78 with JobSpecification

use of org.apache.hyracks.api.job.JobSpecification in project asterixdb by apache.

the class DataflowTest method testHDFSReadWriteOperators.

/**
     * Test a job with only HDFS read and writes.
     *
     * @throws Exception
     */
@SuppressWarnings({ "rawtypes", "unchecked" })
public void testHDFSReadWriteOperators() throws Exception {
    FileInputFormat.setInputPaths(conf, HDFS_INPUT_PATH);
    FileOutputFormat.setOutputPath(conf, new Path(HDFS_OUTPUT_PATH));
    conf.setInputFormatClass(TextInputFormat.class);
    Scheduler scheduler = new Scheduler(HyracksUtils.CC_HOST, HyracksUtils.TEST_HYRACKS_CC_CLIENT_PORT);
    InputFormat inputFormat = ReflectionUtils.newInstance(conf.getInputFormatClass(), getConfiguration());
    List<InputSplit> splits = inputFormat.getSplits(conf);
    String[] readSchedule = scheduler.getLocationConstraints(splits);
    JobSpecification jobSpec = new JobSpecification();
    RecordDescriptor recordDesc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer() });
    String[] locations = new String[] { HyracksUtils.NC1_ID, HyracksUtils.NC1_ID, HyracksUtils.NC2_ID, HyracksUtils.NC2_ID };
    HDFSReadOperatorDescriptor readOperator = new HDFSReadOperatorDescriptor(jobSpec, recordDesc, conf, splits, readSchedule, new TextKeyValueParserFactory());
    PartitionConstraintHelper.addAbsoluteLocationConstraint(jobSpec, readOperator, locations);
    ExternalSortOperatorDescriptor sortOperator = new ExternalSortOperatorDescriptor(jobSpec, 10, new int[] { 0 }, new IBinaryComparatorFactory[] { RawBinaryComparatorFactory.INSTANCE }, recordDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(jobSpec, sortOperator, locations);
    HDFSWriteOperatorDescriptor writeOperator = new HDFSWriteOperatorDescriptor(jobSpec, conf, new TextTupleWriterFactory());
    PartitionConstraintHelper.addAbsoluteLocationConstraint(jobSpec, writeOperator, HyracksUtils.NC1_ID);
    jobSpec.connect(new OneToOneConnectorDescriptor(jobSpec), readOperator, 0, sortOperator, 0);
    jobSpec.connect(new MToNPartitioningMergingConnectorDescriptor(jobSpec, new FieldHashPartitionComputerFactory(new int[] { 0 }, new IBinaryHashFunctionFactory[] { RawBinaryHashFunctionFactory.INSTANCE }), new int[] { 0 }, new IBinaryComparatorFactory[] { RawBinaryComparatorFactory.INSTANCE }, null), sortOperator, 0, writeOperator, 0);
    jobSpec.addRoot(writeOperator);
    IHyracksClientConnection client = new HyracksConnection(HyracksUtils.CC_HOST, HyracksUtils.TEST_HYRACKS_CC_CLIENT_PORT);
    JobId jobId = client.startJob(jobSpec);
    client.waitForCompletion(jobId);
    Assert.assertEquals(true, checkResults());
}

Example 79 with JobSpecification

use of org.apache.hyracks.api.job.JobSpecification in project asterixdb by apache.

the class TPCHCustomerOptimizedHybridHashJoinTest method customerOrderCIDHybridHashJoin_Case1.

@Test
public void customerOrderCIDHybridHashJoin_Case1() throws Exception {
    JobSpecification spec = new JobSpecification();
    FileSplit[] custSplits = new FileSplit[] { new ManagedFileSplit(NC1_ID, "data" + File.separator + "tpch0.001" + File.separator + "customer4.tbl") };
    IFileSplitProvider custSplitsProvider = new ConstantFileSplitProvider(custSplits);
    FileSplit[] ordersSplits = new FileSplit[] { new ManagedFileSplit(NC2_ID, "data" + File.separator + "tpch0.001" + File.separator + "orders4.tbl") };
    IFileSplitProvider ordersSplitsProvider = new ConstantFileSplitProvider(ordersSplits);
    FileScanOperatorDescriptor ordScanner = new FileScanOperatorDescriptor(spec, ordersSplitsProvider, new DelimitedDataTupleParserFactory(orderValueParserFactories, '|'), ordersDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, ordScanner, NC2_ID);
    FileScanOperatorDescriptor custScanner = new FileScanOperatorDescriptor(spec, custSplitsProvider, new DelimitedDataTupleParserFactory(custValueParserFactories, '|'), custDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, custScanner, NC1_ID);
    OptimizedHybridHashJoinOperatorDescriptor join = new OptimizedHybridHashJoinOperatorDescriptor(spec, 15, 243, 1.2, new int[] { 0 }, new int[] { 1 }, new IBinaryHashFunctionFamily[] { UTF8StringBinaryHashFunctionFamily.INSTANCE }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, custOrderJoinDesc, new JoinComparatorFactory(PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), 0, 1), new JoinComparatorFactory(PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), 1, 0), null);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, join, NC1_ID);
    String path = getClass().getName() + File.separator + "case1";
    IOperatorDescriptor printer = getPrinter(spec, path);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
    IConnectorDescriptor custJoinConn = new OneToOneConnectorDescriptor(spec);
    spec.connect(custJoinConn, custScanner, 0, join, 0);
    IConnectorDescriptor ordJoinConn = new MToNBroadcastConnectorDescriptor(spec);
    spec.connect(ordJoinConn, ordScanner, 0, join, 1);
    IConnectorDescriptor joinPrinterConn = new OneToOneConnectorDescriptor(spec);
    spec.connect(joinPrinterConn, join, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
    System.out.println("output to " + path);
}

Example 80 with JobSpecification

use of org.apache.hyracks.api.job.JobSpecification in project asterixdb by apache.

the class TPCHCustomerOrderHashJoinTest method customerOrderCIDJoin.

/*
     * TPCH Customer table: CREATE TABLE CUSTOMER ( C_CUSTKEY INTEGER NOT NULL,
     * C_NAME VARCHAR(25) NOT NULL, C_ADDRESS VARCHAR(40) NOT NULL, C_NATIONKEY
     * INTEGER NOT NULL, C_PHONE CHAR(15) NOT NULL, C_ACCTBAL DECIMAL(15,2) NOT
     * NULL, C_MKTSEGMENT CHAR(10) NOT NULL, C_COMMENT VARCHAR(117) NOT NULL );
     * TPCH Orders table: CREATE TABLE ORDERS ( O_ORDERKEY INTEGER NOT NULL,
     * O_CUSTKEY INTEGER NOT NULL, O_ORDERSTATUS CHAR(1) NOT NULL, O_TOTALPRICE
     * DECIMAL(15,2) NOT NULL, O_ORDERDATE DATE NOT NULL, O_ORDERPRIORITY
     * CHAR(15) NOT NULL, O_CLERK CHAR(15) NOT NULL, O_SHIPPRIORITY INTEGER NOT
     * NULL, O_COMMENT VARCHAR(79) NOT NULL );
     */
@Test
public void customerOrderCIDJoin() throws Exception {
    JobSpecification spec = new JobSpecification();
    FileSplit[] custSplits = new FileSplit[] { new ManagedFileSplit(NC1_ID, "data" + File.separator + "tpch0.001" + File.separator + "customer.tbl") };
    IFileSplitProvider custSplitsProvider = new ConstantFileSplitProvider(custSplits);
    RecordDescriptor custDesc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() });
    FileSplit[] ordersSplits = new FileSplit[] { new ManagedFileSplit(NC2_ID, "data" + File.separator + "tpch0.001" + File.separator + "orders.tbl") };
    IFileSplitProvider ordersSplitsProvider = new ConstantFileSplitProvider(ordersSplits);
    RecordDescriptor ordersDesc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() });
    RecordDescriptor custOrderJoinDesc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() });
    FileScanOperatorDescriptor ordScanner = new FileScanOperatorDescriptor(spec, ordersSplitsProvider, new DelimitedDataTupleParserFactory(new IValueParserFactory[] { UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE }, '|'), ordersDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, ordScanner, NC2_ID);
    FileScanOperatorDescriptor custScanner = new FileScanOperatorDescriptor(spec, custSplitsProvider, new DelimitedDataTupleParserFactory(new IValueParserFactory[] { UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE }, '|'), custDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, custScanner, NC1_ID);
    InMemoryHashJoinOperatorDescriptor join = new InMemoryHashJoinOperatorDescriptor(spec, new int[] { 1 }, new int[] { 0 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, custOrderJoinDesc, 128, null, 128);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, join, NC1_ID);
    ResultSetId rsId = new ResultSetId(1);
    spec.addResultSetId(rsId);
    IOperatorDescriptor printer = new ResultWriterOperatorDescriptor(spec, rsId, false, false, ResultSerializerFactoryProvider.INSTANCE.getResultSerializerFactoryProvider());
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
    IConnectorDescriptor ordJoinConn = new MToNBroadcastConnectorDescriptor(spec);
    spec.connect(ordJoinConn, ordScanner, 0, join, 0);
    IConnectorDescriptor custJoinConn = new OneToOneConnectorDescriptor(spec);
    spec.connect(custJoinConn, custScanner, 0, join, 1);
    IConnectorDescriptor joinPrinterConn = new OneToOneConnectorDescriptor(spec);
    spec.connect(joinPrinterConn, join, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
}