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

Example 81 with IFileSplitProvider

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

the class MetadataProvider method buildRtreeRuntime.

public Pair<IOperatorDescriptor, AlgebricksPartitionConstraint> buildRtreeRuntime(JobSpecification jobSpec, List<LogicalVariable> outputVars, IOperatorSchema opSchema, IVariableTypeEnvironment typeEnv, JobGenContext context, boolean retainInput, boolean retainMissing, Dataset dataset, String indexName, int[] keyFields, int[] minFilterFieldIndexes, int[] maxFilterFieldIndexes) throws AlgebricksException {
    try {
        int numPrimaryKeys = dataset.getPrimaryKeys().size();
        Index secondaryIndex = MetadataManager.INSTANCE.getIndex(mdTxnCtx, dataset.getDataverseName(), dataset.getDatasetName(), indexName);
        if (secondaryIndex == null) {
            throw new AlgebricksException("Code generation error: no index " + indexName + " for dataset " + dataset.getDatasetName());
        }
        RecordDescriptor outputRecDesc = JobGenHelper.mkRecordDescriptor(typeEnv, opSchema, context);
        Pair<IFileSplitProvider, AlgebricksPartitionConstraint> spPc = getSplitProviderAndConstraints(dataset, secondaryIndex.getIndexName());
        int[] primaryKeyFields = new int[numPrimaryKeys];
        for (int i = 0; i < numPrimaryKeys; i++) {
            primaryKeyFields[i] = i;
        }
        ISearchOperationCallbackFactory searchCallbackFactory = dataset.getSearchCallbackFactory(storaegComponentProvider, secondaryIndex, jobId, IndexOperation.SEARCH, primaryKeyFields);
        RTreeSearchOperatorDescriptor rtreeSearchOp;
        IIndexDataflowHelperFactory indexDataflowHelperFactory = new IndexDataflowHelperFactory(storaegComponentProvider.getStorageManager(), spPc.first);
        if (dataset.getDatasetType() == DatasetType.INTERNAL) {
            rtreeSearchOp = new RTreeSearchOperatorDescriptor(jobSpec, outputRecDesc, keyFields, true, true, indexDataflowHelperFactory, retainInput, retainMissing, context.getMissingWriterFactory(), searchCallbackFactory, minFilterFieldIndexes, maxFilterFieldIndexes, false);
        } else {
            // Create the operator
            rtreeSearchOp = new ExternalRTreeSearchOperatorDescriptor(jobSpec, outputRecDesc, keyFields, true, true, indexDataflowHelperFactory, retainInput, retainMissing, context.getMissingWriterFactory(), searchCallbackFactory, minFilterFieldIndexes, maxFilterFieldIndexes, ExternalDatasetsRegistry.INSTANCE.getAndLockDatasetVersion(dataset, this));
        }
        return new Pair<>(rtreeSearchOp, spPc.second);
    } catch (MetadataException me) {
        throw new AlgebricksException(me);
    }
}

Example 82 with IFileSplitProvider

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

the class PushRuntimeTest method scanMicroSortWrite.

@Test
public void scanMicroSortWrite() throws Exception {
    JobSpecification spec = new JobSpecification(FRAME_SIZE);
    // the scanner
    FileSplit[] fileSplits = new FileSplit[1];
    fileSplits[0] = new ManagedFileSplit(AlgebricksHyracksIntegrationUtil.NC1_ID, "data" + File.separator + "tpch0.001" + File.separator + "nation.tbl");
    IFileSplitProvider splitProvider = new ConstantFileSplitProvider(fileSplits);
    RecordDescriptor scannerDesc = new RecordDescriptor(new ISerializerDeserializer[] { IntegerSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer() });
    IValueParserFactory[] valueParsers = new IValueParserFactory[] { IntegerParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, IntegerParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE };
    FileScanOperatorDescriptor scanner = new FileScanOperatorDescriptor(spec, splitProvider, new DelimitedDataTupleParserFactory(valueParsers, '|'), scannerDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, scanner, new String[] { AlgebricksHyracksIntegrationUtil.NC1_ID });
    // the algebricks op.
    InMemorySortRuntimeFactory sort = new InMemorySortRuntimeFactory(new int[] { 1 }, null, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, null);
    RecordDescriptor sortDesc = scannerDesc;
    String fileName = "scanMicroSortWrite.out";
    String filePath = PATH_ACTUAL + SEPARATOR + fileName;
    String resultFilePath = PATH_EXPECTED + SEPARATOR + fileName;
    File outFile = new File(filePath);
    SinkWriterRuntimeFactory writer = new SinkWriterRuntimeFactory(new int[] { 0, 1, 2, 3 }, new IPrinterFactory[] { IntegerPrinterFactory.INSTANCE, UTF8StringPrinterFactory.INSTANCE, IntegerPrinterFactory.INSTANCE, UTF8StringPrinterFactory.INSTANCE }, outFile, PrinterBasedWriterFactory.INSTANCE, sortDesc);
    AlgebricksMetaOperatorDescriptor algebricksOp = new AlgebricksMetaOperatorDescriptor(spec, 1, 0, new IPushRuntimeFactory[] { sort, writer }, new RecordDescriptor[] { sortDesc, null });
    PartitionConstraintHelper.addPartitionCountConstraint(spec, algebricksOp, 1);
    spec.connect(new OneToOneConnectorDescriptor(spec), scanner, 0, algebricksOp, 0);
    spec.addRoot(algebricksOp);
    AlgebricksHyracksIntegrationUtil.runJob(spec);
    compareFiles(filePath, resultFilePath);
    outFile.delete();
}

Example 83 with IFileSplitProvider

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

the class PushRuntimeTest method scanMicroSortGbySelectWrite.

@Test
public void scanMicroSortGbySelectWrite() throws Exception {
    JobSpecification spec = new JobSpecification(FRAME_SIZE);
    // the scanner
    FileSplit[] fileSplits = new FileSplit[1];
    fileSplits[0] = new ManagedFileSplit(AlgebricksHyracksIntegrationUtil.NC1_ID, "data" + File.separator + "tpch0.001" + File.separator + "customer.tbl");
    IFileSplitProvider splitProvider = new ConstantFileSplitProvider(fileSplits);
    RecordDescriptor scannerDesc = new RecordDescriptor(new ISerializerDeserializer[] { IntegerSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer(), FloatSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() });
    IValueParserFactory[] valueParsers = new IValueParserFactory[] { IntegerParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, IntegerParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, FloatParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE };
    FileScanOperatorDescriptor scanner = new FileScanOperatorDescriptor(spec, splitProvider, new DelimitedDataTupleParserFactory(valueParsers, '|'), scannerDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, scanner, new String[] { AlgebricksHyracksIntegrationUtil.NC1_ID });
    // the sort (by nation id)
    RecordDescriptor sortDesc = scannerDesc;
    InMemorySortRuntimeFactory sort = new InMemorySortRuntimeFactory(new int[] { 3 }, null, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY) }, null);
    // the group-by
    NestedTupleSourceRuntimeFactory nts = new NestedTupleSourceRuntimeFactory();
    RecordDescriptor ntsDesc = sortDesc;
    AggregateRuntimeFactory agg = new AggregateRuntimeFactory(new IAggregateEvaluatorFactory[] { new TupleCountAggregateFunctionFactory() });
    RecordDescriptor aggDesc = new RecordDescriptor(new ISerializerDeserializer[] { IntegerSerializerDeserializer.INSTANCE });
    AlgebricksPipeline pipeline = new AlgebricksPipeline(new IPushRuntimeFactory[] { nts, agg }, new RecordDescriptor[] { ntsDesc, aggDesc });
    NestedPlansAccumulatingAggregatorFactory npaaf = new NestedPlansAccumulatingAggregatorFactory(new AlgebricksPipeline[] { pipeline }, new int[] { 3 }, new int[] {});
    RecordDescriptor gbyDesc = new RecordDescriptor(new ISerializerDeserializer[] { IntegerSerializerDeserializer.INSTANCE, IntegerSerializerDeserializer.INSTANCE });
    MicroPreClusteredGroupRuntimeFactory gby = new MicroPreClusteredGroupRuntimeFactory(new int[] { 3 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY) }, npaaf, sortDesc, gbyDesc, null);
    // the algebricks op.
    IScalarEvaluatorFactory cond = new IntegerEqualsEvalFactory(new IntegerConstantEvalFactory(3), // Canadian customers
    new TupleFieldEvaluatorFactory(0));
    StreamSelectRuntimeFactory select = new StreamSelectRuntimeFactory(cond, new int[] { 1 }, BinaryBooleanInspectorImpl.FACTORY, false, -1, null);
    RecordDescriptor selectDesc = new RecordDescriptor(new ISerializerDeserializer[] { IntegerSerializerDeserializer.INSTANCE });
    String filePath = PATH_ACTUAL + SEPARATOR + "scanSortGbySelectWrite.out";
    File outFile = new File(filePath);
    SinkWriterRuntimeFactory writer = new SinkWriterRuntimeFactory(new int[] { 0 }, new IPrinterFactory[] { IntegerPrinterFactory.INSTANCE }, outFile, PrinterBasedWriterFactory.INSTANCE, selectDesc);
    AlgebricksMetaOperatorDescriptor algebricksOp = new AlgebricksMetaOperatorDescriptor(spec, 1, 0, new IPushRuntimeFactory[] { sort, gby, select, writer }, new RecordDescriptor[] { sortDesc, gbyDesc, selectDesc, null });
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, algebricksOp, new String[] { AlgebricksHyracksIntegrationUtil.NC1_ID });
    spec.connect(new OneToOneConnectorDescriptor(spec), scanner, 0, algebricksOp, 0);
    spec.addRoot(algebricksOp);
    AlgebricksHyracksIntegrationUtil.runJob(spec);
    StringBuilder buf = new StringBuilder();
    readFileToString(outFile, buf);
    Assert.assertEquals("9", buf.toString());
    outFile.delete();
}

Example 84 with IFileSplitProvider

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

the class PushRuntimeTest method scanAggregateWrite.

@Test
public void scanAggregateWrite() throws Exception {
    JobSpecification spec = new JobSpecification(FRAME_SIZE);
    // the scanner
    FileSplit[] fileSplits = new FileSplit[1];
    fileSplits[0] = new ManagedFileSplit(AlgebricksHyracksIntegrationUtil.NC1_ID, "data" + File.separator + "tpch0.001" + File.separator + "customer-part1.tbl");
    IFileSplitProvider splitProvider = new ConstantFileSplitProvider(fileSplits);
    RecordDescriptor scannerDesc = new RecordDescriptor(new ISerializerDeserializer[] { IntegerSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer(), FloatSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() });
    IValueParserFactory[] valueParsers = new IValueParserFactory[] { IntegerParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, IntegerParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, FloatParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE, UTF8StringParserFactory.INSTANCE };
    FileScanOperatorDescriptor scanner = new FileScanOperatorDescriptor(spec, splitProvider, new DelimitedDataTupleParserFactory(valueParsers, '|'), scannerDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, scanner, new String[] { AlgebricksHyracksIntegrationUtil.NC1_ID });
    // the algebricks op.
    AggregateRuntimeFactory agg = new AggregateRuntimeFactory(new IAggregateEvaluatorFactory[] { new TupleCountAggregateFunctionFactory() });
    RecordDescriptor aggDesc = new RecordDescriptor(new ISerializerDeserializer[] { IntegerSerializerDeserializer.INSTANCE });
    String filePath = PATH_ACTUAL + SEPARATOR + "scanAggregateWrite.out";
    File outFile = new File(filePath);
    SinkWriterRuntimeFactory writer = new SinkWriterRuntimeFactory(new int[] { 0 }, new IPrinterFactory[] { IntegerPrinterFactory.INSTANCE }, outFile, PrinterBasedWriterFactory.INSTANCE, aggDesc);
    AlgebricksMetaOperatorDescriptor algebricksOp = new AlgebricksMetaOperatorDescriptor(spec, 1, 0, new IPushRuntimeFactory[] { agg, writer }, new RecordDescriptor[] { aggDesc, null });
    PartitionConstraintHelper.addPartitionCountConstraint(spec, algebricksOp, 1);
    spec.connect(new OneToOneConnectorDescriptor(spec), scanner, 0, algebricksOp, 0);
    spec.addRoot(algebricksOp);
    AlgebricksHyracksIntegrationUtil.runJob(spec);
    StringBuilder buf = new StringBuilder();
    readFileToString(outFile, buf);
    Assert.assertEquals("75", buf.toString());
    outFile.delete();
}

Example 85 with IFileSplitProvider

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

the class VSizeFrameSortMergeTest method sortTask.

public void sortTask(int frameSize, int frameLimit) throws Exception {
    JobSpecification spec = new JobSpecification();
    FileScanOperatorDescriptor ordScanner = new FileScanOperatorDescriptor(spec, ordersSplitProvider, 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, NC1_ID, NC2_ID);
    spec.setFrameSize(frameSize);
    ExternalSortOperatorDescriptor sorter = new ExternalSortOperatorDescriptor(spec, frameLimit, new int[] { 1, 0 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, ordersDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, sorter, NC1_ID, NC2_ID);
    String path = getClass().getSimpleName() + aInteger.getAndIncrement() + ".tmp";
    IFileSplitProvider outputSplitProvider = new ConstantFileSplitProvider(new FileSplit[] { new ManagedFileSplit(NC1_ID, path) });
    IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outputSplitProvider, "|");
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
    spec.connect(new OneToOneConnectorDescriptor(spec), ordScanner, 0, sorter, 0);
    spec.connect(new MToNPartitioningMergingConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(new int[] { 1, 0 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY), PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }), new int[] { 1, 0 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new UTF8StringNormalizedKeyComputerFactory()), sorter, 0, printer, 0);
    spec.addRoot(printer);
    runTest(spec);
    System.out.println("Result write into :" + path);
}