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

Example 51 with IFileSplitProvider

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

the class UnionTest method createUnionJobSpec.

public static JobSpecification createUnionJobSpec() throws Exception {
    JobSpecification spec = new JobSpecification();
    IFileSplitProvider splitProvider = new ConstantFileSplitProvider(new FileSplit[] { new ManagedFileSplit(NC2_ID, "data" + File.separator + "words.txt"), new ManagedFileSplit(NC1_ID, "data" + File.separator + "nc1" + File.separator + "words.txt") });
    RecordDescriptor desc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer() });
    FileScanOperatorDescriptor csvScanner01 = new FileScanOperatorDescriptor(spec, splitProvider, new DelimitedDataTupleParserFactory(new IValueParserFactory[] { UTF8StringParserFactory.INSTANCE }, ','), desc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, csvScanner01, NC2_ID, NC1_ID);
    FileScanOperatorDescriptor csvScanner02 = new FileScanOperatorDescriptor(spec, splitProvider, new DelimitedDataTupleParserFactory(new IValueParserFactory[] { UTF8StringParserFactory.INSTANCE }, ','), desc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, csvScanner02, NC2_ID, NC1_ID);
    UnionAllOperatorDescriptor unionAll = new UnionAllOperatorDescriptor(spec, 2, desc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, unionAll, NC2_ID, 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, NC2_ID, NC1_ID);
    spec.connect(new OneToOneConnectorDescriptor(spec), csvScanner01, 0, unionAll, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), csvScanner02, 0, unionAll, 1);
    spec.connect(new OneToOneConnectorDescriptor(spec), unionAll, 0, printer, 0);
    spec.addRoot(printer);
    return spec;
}

Example 52 with IFileSplitProvider

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

the class DataverseUtil method dropDataverseJobSpec.

public static JobSpecification dropDataverseJobSpec(Dataverse dataverse, MetadataProvider metadata) {
    JobSpecification jobSpec = RuntimeUtils.createJobSpecification(metadata.getApplicationContext());
    Pair<IFileSplitProvider, AlgebricksPartitionConstraint> splitsAndConstraint = metadata.splitAndConstraints(dataverse.getDataverseName());
    FileRemoveOperatorDescriptor frod = new FileRemoveOperatorDescriptor(jobSpec, splitsAndConstraint.first, false);
    AlgebricksPartitionConstraintHelper.setPartitionConstraintInJobSpec(jobSpec, frod, splitsAndConstraint.second);
    jobSpec.addRoot(frod);
    return jobSpec;
}

Example 53 with IFileSplitProvider

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

the class FlushDatasetUtil method flushDataset.

public static void flushDataset(IHyracksClientConnection hcc, MetadataProvider metadataProvider, String dataverseName, String datasetName, String indexName) throws Exception {
    CompilerProperties compilerProperties = metadataProvider.getApplicationContext().getCompilerProperties();
    int frameSize = compilerProperties.getFrameSize();
    JobSpecification spec = new JobSpecification(frameSize);
    RecordDescriptor[] rDescs = new RecordDescriptor[] { new RecordDescriptor(new ISerializerDeserializer[] {}) };
    AlgebricksMetaOperatorDescriptor emptySource = new AlgebricksMetaOperatorDescriptor(spec, 0, 1, new IPushRuntimeFactory[] { new EmptyTupleSourceRuntimeFactory() }, rDescs);
    org.apache.asterix.common.transactions.JobId jobId = JobIdFactory.generateJobId();
    Dataset dataset = metadataProvider.findDataset(dataverseName, datasetName);
    FlushDatasetOperatorDescriptor flushOperator = new FlushDatasetOperatorDescriptor(spec, jobId, dataset.getDatasetId());
    spec.connect(new OneToOneConnectorDescriptor(spec), emptySource, 0, flushOperator, 0);
    Pair<IFileSplitProvider, AlgebricksPartitionConstraint> primarySplitsAndConstraint = metadataProvider.getSplitProviderAndConstraints(dataset, indexName);
    AlgebricksPartitionConstraint primaryPartitionConstraint = primarySplitsAndConstraint.second;
    AlgebricksPartitionConstraintHelper.setPartitionConstraintInJobSpec(spec, emptySource, primaryPartitionConstraint);
    JobEventListenerFactory jobEventListenerFactory = new JobEventListenerFactory(jobId, true);
    spec.setJobletEventListenerFactory(jobEventListenerFactory);
    JobUtils.runJob(hcc, spec, true);
}

Example 54 with IFileSplitProvider

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

the class SecondaryTreeIndexOperationsHelper method buildDropJobSpec.

@Override
public JobSpecification buildDropJobSpec() throws AlgebricksException {
    JobSpecification spec = RuntimeUtils.createJobSpecification(metadataProvider.getApplicationContext());
    Pair<IFileSplitProvider, AlgebricksPartitionConstraint> splitsAndConstraint = metadataProvider.getSplitProviderAndConstraints(dataset, index.getIndexName());
    IIndexDataflowHelperFactory dataflowHelperFactory = new IndexDataflowHelperFactory(metadataProvider.getStorageComponentProvider().getStorageManager(), splitsAndConstraint.first);
    // The index drop operation should be persistent regardless of temp datasets or permanent dataset.
    IndexDropOperatorDescriptor btreeDrop = new IndexDropOperatorDescriptor(spec, dataflowHelperFactory);
    AlgebricksPartitionConstraintHelper.setPartitionConstraintInJobSpec(spec, btreeDrop, splitsAndConstraint.second);
    spec.addRoot(btreeDrop);
    return spec;
}

Example 55 with IFileSplitProvider

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

the class SecondaryIndexBulkLoadExample method createJob.

private static JobSpecification createJob(Options options) {
    JobSpecification spec = new JobSpecification(options.frameSize);
    String[] splitNCs = options.ncs.split(",");
    IStorageManager storageManager = BTreeHelperStorageManager.INSTANCE;
    // schema of tuples that we are retrieving from the primary index
    RecordDescriptor recDesc = new RecordDescriptor(new ISerializerDeserializer[] { // we will use this as payload in secondary index
    IntegerSerializerDeserializer.INSTANCE, // we will use this ask key in secondary index
    new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer() });
    int primaryFieldCount = 4;
    ITypeTraits[] primaryTypeTraits = new ITypeTraits[primaryFieldCount];
    primaryTypeTraits[0] = IntegerPointable.TYPE_TRAITS;
    primaryTypeTraits[1] = UTF8StringPointable.TYPE_TRAITS;
    primaryTypeTraits[2] = IntegerPointable.TYPE_TRAITS;
    primaryTypeTraits[3] = UTF8StringPointable.TYPE_TRAITS;
    // comparators for sort fields and BTree fields
    IBinaryComparatorFactory[] comparatorFactories = new IBinaryComparatorFactory[2];
    comparatorFactories[0] = PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY);
    comparatorFactories[1] = PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY);
    // use a disk-order scan to read primary index
    IFileSplitProvider primarySplitProvider = JobHelper.createFileSplitProvider(splitNCs, options.primaryBTreeName);
    IIndexDataflowHelperFactory primaryHelperFactory = new IndexDataflowHelperFactory(storageManager, primarySplitProvider);
    TreeIndexDiskOrderScanOperatorDescriptor btreeScanOp = new TreeIndexDiskOrderScanOperatorDescriptor(spec, recDesc, primaryHelperFactory, NoOpOperationCallbackFactory.INSTANCE);
    JobHelper.createPartitionConstraint(spec, btreeScanOp, splitNCs);
    // sort the tuples as preparation for bulk load into secondary index
    // fields to sort on
    int[] sortFields = { 1, 0 };
    ExternalSortOperatorDescriptor sorter = new ExternalSortOperatorDescriptor(spec, options.sbSize, sortFields, comparatorFactories, recDesc);
    JobHelper.createPartitionConstraint(spec, sorter, splitNCs);
    // tuples to be put into B-Tree shall have 2 fields
    int secondaryFieldCount = 2;
    ITypeTraits[] secondaryTypeTraits = new ITypeTraits[secondaryFieldCount];
    secondaryTypeTraits[0] = UTF8StringPointable.TYPE_TRAITS;
    secondaryTypeTraits[1] = IntegerPointable.TYPE_TRAITS;
    // the B-Tree expects its keyfields to be at the front of its input
    // tuple
    int[] fieldPermutation = { 1, 0 };
    IFileSplitProvider btreeSplitProvider = JobHelper.createFileSplitProvider(splitNCs, options.secondaryBTreeName);
    IIndexDataflowHelperFactory secondaryHelperFactory = new IndexDataflowHelperFactory(storageManager, btreeSplitProvider);
    TreeIndexBulkLoadOperatorDescriptor btreeBulkLoad = new TreeIndexBulkLoadOperatorDescriptor(spec, null, fieldPermutation, 0.7f, false, 1000L, true, secondaryHelperFactory);
    JobHelper.createPartitionConstraint(spec, btreeBulkLoad, splitNCs);
    NullSinkOperatorDescriptor nsOpDesc = new NullSinkOperatorDescriptor(spec);
    JobHelper.createPartitionConstraint(spec, nsOpDesc, splitNCs);
    // connect the ops
    spec.connect(new OneToOneConnectorDescriptor(spec), btreeScanOp, 0, sorter, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), sorter, 0, btreeBulkLoad, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), btreeBulkLoad, 0, nsOpDesc, 0);
    spec.addRoot(nsOpDesc);
    return spec;
}