Examples with RecordDescriptor org.apache.hyracks.api.dataflow.value.RecordDescriptor used on opensource projects

Example 71 with RecordDescriptor

use of org.apache.hyracks.api.dataflow.value.RecordDescriptor in project asterixdb by apache.

the class HDFSWriteOperatorDescriptor method createPushRuntime.

@Override
public IOperatorNodePushable createPushRuntime(final IHyracksTaskContext ctx, final IRecordDescriptorProvider recordDescProvider, final int partition, final int nPartitions) throws HyracksDataException {
    return new AbstractUnaryInputSinkOperatorNodePushable() {

        private FSDataOutputStream dos;

        private RecordDescriptor inputRd = recordDescProvider.getInputRecordDescriptor(getActivityId(), 0);

        ;

        private FrameTupleAccessor accessor = new FrameTupleAccessor(inputRd);

        private FrameTupleReference tuple = new FrameTupleReference();

        private ITupleWriter tupleWriter;

        private ClassLoader ctxCL;

        @Override
        public void open() throws HyracksDataException {
            ctxCL = Thread.currentThread().getContextClassLoader();
            Thread.currentThread().setContextClassLoader(this.getClass().getClassLoader());
            JobConf conf = confFactory.getConf();
            String outputDirPath = FileOutputFormat.getOutputPath(conf).toString();
            String fileName = outputDirPath + File.separator + "part-" + partition;
            tupleWriter = tupleWriterFactory.getTupleWriter(ctx, partition, nPartitions);
            try {
                FileSystem dfs = FileSystem.get(conf);
                dos = dfs.create(new Path(fileName), true);
                tupleWriter.open(dos);
            } catch (Exception e) {
                throw new HyracksDataException(e);
            }
        }

        @Override
        public void nextFrame(ByteBuffer buffer) throws HyracksDataException {
            accessor.reset(buffer);
            int tupleCount = accessor.getTupleCount();
            for (int i = 0; i < tupleCount; i++) {
                tuple.reset(accessor, i);
                tupleWriter.write(dos, tuple);
            }
        }

        @Override
        public void fail() throws HyracksDataException {
        }

        @Override
        public void close() throws HyracksDataException {
            try {
                tupleWriter.close(dos);
                dos.close();
            } catch (Exception e) {
                throw new HyracksDataException(e);
            } finally {
                Thread.currentThread().setContextClassLoader(ctxCL);
            }
        }
    };
}

Example 72 with RecordDescriptor

use of org.apache.hyracks.api.dataflow.value.RecordDescriptor in project asterixdb by apache.

the class ConnectorDescriptorWithMessagingTest method testEmptyFrames.

@Test
public void testEmptyFrames() throws Exception {
    try {
        List<Integer> routing = Arrays.asList(0, 1, 2, 3, 4);
        IConnectorDescriptorRegistry connDescRegistry = Mockito.mock(IConnectorDescriptorRegistry.class);
        ITuplePartitionComputerFactory partitionComputerFactory = new TestPartitionComputerFactory(routing);
        MToNPartitioningWithMessageConnectorDescriptor connector = new MToNPartitioningWithMessageConnectorDescriptor(connDescRegistry, partitionComputerFactory);
        IHyracksTaskContext ctx = TestUtils.create(DEFAULT_FRAME_SIZE);
        VSizeFrame message = new VSizeFrame(ctx);
        VSizeFrame tempBuffer = new VSizeFrame(ctx);
        TaskUtil.putInSharedMap(HyracksConstants.KEY_MESSAGE, message, ctx);
        message.getBuffer().clear();
        message.getBuffer().put(MessagingFrameTupleAppender.NULL_FEED_MESSAGE);
        message.getBuffer().flip();
        ISerializerDeserializer<?>[] serdes = new ISerializerDeserializer<?>[] { Integer64SerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE, BooleanSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer() };
        RecordDescriptor rDesc = new RecordDescriptor(serdes);
        TestPartitionWriterFactory partitionWriterFactory = new TestPartitionWriterFactory();
        IFrameWriter partitioner = connector.createPartitioner(ctx, rDesc, partitionWriterFactory, CURRENT_PRODUCER, NUMBER_OF_CONSUMERS, NUMBER_OF_CONSUMERS);
        List<TestFrameWriter> recipients = new ArrayList<>();
        try {
            partitioner.open();
            FrameTupleAccessor fta = new FrameTupleAccessor(rDesc);
            for (IFrameWriter writer : partitionWriterFactory.getWriters().values()) {
                recipients.add((TestFrameWriter) writer);
            }
            partitioner.flush();
            for (TestFrameWriter writer : recipients) {
                Assert.assertEquals(writer.nextFrameCount(), 1);
                fta.reset(writer.getLastFrame());
                Assert.assertEquals(fta.getTupleCount(), 1);
                FeedUtils.processFeedMessage(writer.getLastFrame(), tempBuffer, fta);
                Assert.assertEquals(MessagingFrameTupleAppender.NULL_FEED_MESSAGE, MessagingFrameTupleAppender.getMessageType(tempBuffer));
            }
            message.getBuffer().clear();
            message.getBuffer().put(MessagingFrameTupleAppender.ACK_REQ_FEED_MESSAGE);
            message.getBuffer().flip();
            partitioner.flush();
            for (TestFrameWriter writer : recipients) {
                Assert.assertEquals(writer.nextFrameCount(), 2);
                fta.reset(writer.getLastFrame());
                Assert.assertEquals(fta.getTupleCount(), 1);
                FeedUtils.processFeedMessage(writer.getLastFrame(), tempBuffer, fta);
                Assert.assertEquals(MessagingFrameTupleAppender.ACK_REQ_FEED_MESSAGE, MessagingFrameTupleAppender.getMessageType(tempBuffer));
            }
            message.getBuffer().clear();
            message.getBuffer().put(MessagingFrameTupleAppender.NULL_FEED_MESSAGE);
            message.getBuffer().flip();
            partitioner.flush();
            for (TestFrameWriter writer : recipients) {
                Assert.assertEquals(writer.nextFrameCount(), 3);
                fta.reset(writer.getLastFrame());
                Assert.assertEquals(fta.getTupleCount(), 1);
                FeedUtils.processFeedMessage(writer.getLastFrame(), tempBuffer, fta);
                Assert.assertEquals(MessagingFrameTupleAppender.NULL_FEED_MESSAGE, MessagingFrameTupleAppender.getMessageType(tempBuffer));
            }
        } catch (Throwable t) {
            partitioner.fail();
            throw t;
        } finally {
            partitioner.close();
        }
        for (TestFrameWriter writer : recipients) {
            Assert.assertEquals(writer.nextFrameCount(), 4);
            Assert.assertEquals(writer.closeCount(), 1);
        }
    } catch (Throwable th) {
        th.printStackTrace();
        throw th;
    }
}

Example 73 with RecordDescriptor

use of org.apache.hyracks.api.dataflow.value.RecordDescriptor in project asterixdb by apache.

the class ConnectorDescriptorWithMessagingTest method testMessageFitsWithTuples.

@Test
public void testMessageFitsWithTuples() throws Exception {
    try {
        // Routing will be round robin
        List<Integer> routing = Arrays.asList(0, 1, 2, 3, 4);
        IConnectorDescriptorRegistry connDescRegistry = Mockito.mock(IConnectorDescriptorRegistry.class);
        ITuplePartitionComputerFactory partitionComputerFactory = new TestPartitionComputerFactory(routing);
        MToNPartitioningWithMessageConnectorDescriptor connector = new MToNPartitioningWithMessageConnectorDescriptor(connDescRegistry, partitionComputerFactory);
        IHyracksTaskContext ctx = TestUtils.create(DEFAULT_FRAME_SIZE);
        VSizeFrame message = new VSizeFrame(ctx);
        VSizeFrame tempBuffer = new VSizeFrame(ctx);
        TaskUtil.putInSharedMap(HyracksConstants.KEY_MESSAGE, message, ctx);
        message.getBuffer().clear();
        message.getBuffer().put(MessagingFrameTupleAppender.ACK_REQ_FEED_MESSAGE);
        message.getBuffer().flip();
        ISerializerDeserializer<?>[] serdes = new ISerializerDeserializer<?>[] { Integer64SerializerDeserializer.INSTANCE, DoubleSerializerDeserializer.INSTANCE, BooleanSerializerDeserializer.INSTANCE, new UTF8StringSerializerDeserializer() };
        FieldType[] types = { FieldType.Integer64, FieldType.Double, FieldType.Boolean, FieldType.String };
        RecordDescriptor rDesc = new RecordDescriptor(serdes);
        TestPartitionWriterFactory partitionWriterFactory = new TestPartitionWriterFactory();
        PartitionWithMessageDataWriter partitioner = (PartitionWithMessageDataWriter) connector.createPartitioner(ctx, rDesc, partitionWriterFactory, CURRENT_PRODUCER, NUMBER_OF_CONSUMERS, NUMBER_OF_CONSUMERS);
        partitioner.open();
        FrameTupleAccessor fta = new FrameTupleAccessor(rDesc);
        List<TestFrameWriter> recipients = new ArrayList<>();
        for (int i = 0; i < partitionWriterFactory.getWriters().values().size(); i++) {
            recipients.add(partitionWriterFactory.getWriters().get(i));
        }
        TestTupleGenerator ttg = new TestTupleGenerator(types, STRING_FIELD_SIZES, true);
        VSizeFrame frame = new VSizeFrame(ctx);
        FrameTupleAppender appender = new FrameTupleAppender(frame);
        for (int count = 0; count < NUMBER_OF_CONSUMERS; count++) {
            ITupleReference tuple = ttg.next();
            appender.append(tuple);
        }
        partitioner.nextFrame(frame.getBuffer());
        partitioner.flush();
        Assert.assertEquals(partitionWriterFactory.getWriters().get(0).nextFrameCount(), 1);
        Assert.assertEquals(partitionWriterFactory.getWriters().get(1).nextFrameCount(), 1);
        Assert.assertEquals(partitionWriterFactory.getWriters().get(2).nextFrameCount(), 1);
        Assert.assertEquals(partitionWriterFactory.getWriters().get(3).nextFrameCount(), 1);
        Assert.assertEquals(partitionWriterFactory.getWriters().get(4).nextFrameCount(), 1);
        for (TestFrameWriter writer : recipients) {
            fta.reset(writer.getLastFrame());
            Assert.assertEquals(fta.getTupleCount(), 2);
            FeedUtils.processFeedMessage(writer.getLastFrame(), tempBuffer, fta);
            Assert.assertEquals(MessagingFrameTupleAppender.ACK_REQ_FEED_MESSAGE, MessagingFrameTupleAppender.getMessageType(tempBuffer));
        }
        partitioner.close();
    } catch (Throwable th) {
        th.printStackTrace();
        throw th;
    }
}

Example 74 with RecordDescriptor

use of org.apache.hyracks.api.dataflow.value.RecordDescriptor in project asterixdb by apache.

the class AbstractExternalDatasetIndexesOperatorDescriptor method createPushRuntime.

@Override
public IOperatorNodePushable createPushRuntime(final IHyracksTaskContext ctx, IRecordDescriptorProvider recordDescProvider, final int partition, int nPartitions) {
    return new AbstractOperatorNodePushable() {

        @Override
        public void initialize() throws HyracksDataException {
            try {
                // perform operation on btrees
                for (int i = 0; i < treeIndexesDataflowHelperFactories.size(); i++) {
                    IIndexDataflowHelper indexHelper = treeIndexesDataflowHelperFactories.get(i).create(ctx, partition);
                    performOpOnIndex(indexHelper, ctx);
                }
            } catch (Exception e) {
                throw new HyracksDataException(e);
            }
        }

        @Override
        public void deinitialize() throws HyracksDataException {
        }

        @Override
        public int getInputArity() {
            return 0;
        }

        @Override
        public void setOutputFrameWriter(int index, IFrameWriter writer, RecordDescriptor recordDesc) throws HyracksDataException {
        }

        @Override
        public IFrameWriter getInputFrameWriter(int index) {
            return null;
        }
    };
}

Example 75 with RecordDescriptor

use of org.apache.hyracks.api.dataflow.value.RecordDescriptor 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);
}