Examples with FileSplit org.apache.hyracks.api.io.FileSplit used on opensource projects

Example 46 with FileSplit

use of org.apache.hyracks.api.io.FileSplit in project asterixdb by apache.

the class Common method parseFileSplits.

static FileSplit[] parseFileSplits(String fileSplits) {
    String[] splits = fileSplits.split(",");
    FileSplit[] fSplits = new FileSplit[splits.length];
    for (int i = 0; i < splits.length; ++i) {
        String s = splits[i].trim();
        int idx = s.indexOf(':');
        if (idx < 0) {
            throw new IllegalArgumentException("File split " + s + " not well formed");
        }
        fSplits[i] = new UnmanagedFileSplit(s.substring(0, idx), new File(s.substring(idx + 1)).getAbsolutePath());
    }
    return fSplits;
}

Example 47 with FileSplit

use of org.apache.hyracks.api.io.FileSplit 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 48 with FileSplit

use of org.apache.hyracks.api.io.FileSplit in project asterixdb by apache.

the class AbstractRTreeOperatorTest method loadPrimaryIndex.

protected void loadPrimaryIndex() throws Exception {
    JobSpecification spec = new JobSpecification();
    FileSplit[] ordersSplits = new FileSplit[] { new ManagedFileSplit(NC1_ID, "data" + File.separator + "orders-with-locations-part1.txt") };
    IFileSplitProvider ordersSplitProvider = 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(), DoubleSerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE });
    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, DoubleParserFactory.INSTANCE, DoubleParserFactory.INSTANCE, DoubleParserFactory.INSTANCE, DoubleParserFactory.INSTANCE }, '|'), ordersDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, ordScanner, NC1_ID);
    ExternalSortOperatorDescriptor sorter = new ExternalSortOperatorDescriptor(spec, 1000, new int[] { 0 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, ordersDesc);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, sorter, NC1_ID);
    int[] fieldPermutation = { 0, 1, 2, 4, 5, 7, 9, 10, 11, 12 };
    TreeIndexBulkLoadOperatorDescriptor primaryBulkLoad = new TreeIndexBulkLoadOperatorDescriptor(spec, primaryRecDesc, fieldPermutation, 0.7f, false, 1000L, true, primaryHelperFactory);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, primaryBulkLoad, NC1_ID);
    NullSinkOperatorDescriptor nsOpDesc = new NullSinkOperatorDescriptor(spec);
    PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, nsOpDesc, NC1_ID);
    spec.connect(new OneToOneConnectorDescriptor(spec), ordScanner, 0, sorter, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), sorter, 0, primaryBulkLoad, 0);
    spec.connect(new OneToOneConnectorDescriptor(spec), primaryBulkLoad, 0, nsOpDesc, 0);
    spec.addRoot(nsOpDesc);
    runTest(spec);
}

Example 49 with FileSplit

use of org.apache.hyracks.api.io.FileSplit in project asterixdb by apache.

the class DatasetUtil method createDatasetJobSpec.

public static JobSpecification createDatasetJobSpec(Dataset dataset, MetadataProvider metadataProvider) throws AlgebricksException {
    Index index = IndexUtil.getPrimaryIndex(dataset);
    ARecordType itemType = (ARecordType) metadataProvider.findType(dataset);
    // get meta item type
    ARecordType metaItemType = null;
    if (dataset.hasMetaPart()) {
        metaItemType = (ARecordType) metadataProvider.findMetaType(dataset);
    }
    JobSpecification spec = RuntimeUtils.createJobSpecification(metadataProvider.getApplicationContext());
    Pair<IFileSplitProvider, AlgebricksPartitionConstraint> splitsAndConstraint = metadataProvider.getSplitProviderAndConstraints(dataset);
    FileSplit[] fs = splitsAndConstraint.first.getFileSplits();
    StringBuilder sb = new StringBuilder();
    for (int i = 0; i < fs.length; i++) {
        sb.append(fs[i] + " ");
    }
    LOGGER.info("CREATING File Splits: " + sb.toString());
    Pair<ILSMMergePolicyFactory, Map<String, String>> compactionInfo = DatasetUtil.getMergePolicyFactory(dataset, metadataProvider.getMetadataTxnContext());
    //prepare a LocalResourceMetadata which will be stored in NC's local resource repository
    IResourceFactory resourceFactory = dataset.getResourceFactory(metadataProvider, index, itemType, metaItemType, compactionInfo.first, compactionInfo.second);
    IndexBuilderFactory indexBuilderFactory = new IndexBuilderFactory(metadataProvider.getStorageComponentProvider().getStorageManager(), splitsAndConstraint.first, resourceFactory, !dataset.isTemp());
    IndexCreateOperatorDescriptor indexCreateOp = new IndexCreateOperatorDescriptor(spec, indexBuilderFactory);
    AlgebricksPartitionConstraintHelper.setPartitionConstraintInJobSpec(spec, indexCreateOp, splitsAndConstraint.second);
    spec.addRoot(indexCreateOp);
    return spec;
}

Example 50 with FileSplit

use of org.apache.hyracks.api.io.FileSplit in project asterixdb by apache.

the class PlainFileWriterOperatorDescriptor method createPushRuntime.

/*
     * (non-Javadoc)
     *
     * @see
     * org.apache.hyracks.api.dataflow.IActivityNode#createPushRuntime(edu.
     * uci.ics.hyracks.api.context.IHyracksContext,
     * org.apache.hyracks.api.job.IOperatorEnvironment,
     * org.apache.hyracks.api.dataflow.value.IRecordDescriptorProvider, int,
     * int)
     */
@Override
public IOperatorNodePushable createPushRuntime(IHyracksTaskContext ctx, IRecordDescriptorProvider recordDescProvider, final int partition, int nPartitions) throws HyracksDataException {
    // Output files
    final FileSplit[] splits = fileSplitProvider.getFileSplits();
    IIOManager ioManager = ctx.getIoManager();
    // Frame accessor
    final FrameTupleAccessor frameTupleAccessor = new FrameTupleAccessor(recordDescProvider.getInputRecordDescriptor(getActivityId(), 0));
    // Record descriptor
    final RecordDescriptor recordDescriptor = recordDescProvider.getInputRecordDescriptor(getActivityId(), 0);
    return new AbstractUnaryInputSinkOperatorNodePushable() {

        private BufferedWriter out;

        private ByteBufferInputStream bbis;

        private DataInputStream di;

        @Override
        public void open() throws HyracksDataException {
            try {
                out = new BufferedWriter(new FileWriter(splits[partition].getFile(ioManager)));
                bbis = new ByteBufferInputStream();
                di = new DataInputStream(bbis);
            } catch (Exception e) {
                throw new HyracksDataException(e);
            }
        }

        @Override
        public void nextFrame(ByteBuffer buffer) throws HyracksDataException {
            try {
                frameTupleAccessor.reset(buffer);
                for (int tIndex = 0; tIndex < frameTupleAccessor.getTupleCount(); tIndex++) {
                    int start = frameTupleAccessor.getTupleStartOffset(tIndex) + frameTupleAccessor.getFieldSlotsLength();
                    bbis.setByteBuffer(buffer, start);
                    Object[] record = new Object[recordDescriptor.getFieldCount()];
                    for (int i = 0; i < record.length; ++i) {
                        Object instance = recordDescriptor.getFields()[i].deserialize(di);
                        if (i == 0) {
                            out.write(String.valueOf(instance));
                        } else {
                            out.write(delim + String.valueOf(instance));
                        }
                    }
                    out.write("\n");
                }
            } catch (IOException ex) {
                throw new HyracksDataException(ex);
            }
        }

        @Override
        public void fail() throws HyracksDataException {
        }

        @Override
        public void close() throws HyracksDataException {
            try {
                out.close();
            } catch (IOException e) {
                throw new HyracksDataException(e);
            }
        }
    };
}