Search in sources :

Example 26 with IBinaryComparatorFactory

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

the class OrderedIndexExamplesTest method fixedLengthKeyValueExample.

/**
     * Fixed-Length Key,Value Example. Create a tree index with one fixed-length
     * key field and one fixed-length value field. Fill index with random values
     * using insertions (not bulk load). Perform scans and range search.
     */
@Test
public void fixedLengthKeyValueExample() throws Exception {
    if (LOGGER.isLoggable(Level.INFO)) {
        LOGGER.info("Fixed-Length Key,Value Example.");
    }
    // Declare fields.
    int fieldCount = 2;
    ITypeTraits[] typeTraits = new ITypeTraits[fieldCount];
    typeTraits[0] = IntegerPointable.TYPE_TRAITS;
    typeTraits[1] = IntegerPointable.TYPE_TRAITS;
    // Declare field serdes.
    ISerializerDeserializer[] fieldSerdes = { IntegerSerializerDeserializer.INSTANCE, IntegerSerializerDeserializer.INSTANCE };
    // Declare keys.
    int keyFieldCount = 1;
    IBinaryComparatorFactory[] cmpFactories = new IBinaryComparatorFactory[keyFieldCount];
    cmpFactories[0] = PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY);
    // This is only used for the LSM-BTree.
    int[] bloomFilterKeyFields = new int[keyFieldCount];
    bloomFilterKeyFields[0] = 0;
    ITreeIndex treeIndex = createTreeIndex(typeTraits, cmpFactories, bloomFilterKeyFields, null, null, null, null);
    treeIndex.create();
    treeIndex.activate();
    long start = System.currentTimeMillis();
    if (LOGGER.isLoggable(Level.INFO)) {
        LOGGER.info("Inserting into tree...");
    }
    ArrayTupleBuilder tb = new ArrayTupleBuilder(fieldCount);
    ArrayTupleReference tuple = new ArrayTupleReference();
    IIndexAccessor indexAccessor = treeIndex.createAccessor(TestOperationCallback.INSTANCE, TestOperationCallback.INSTANCE);
    int numInserts = 10000;
    for (int i = 0; i < numInserts; i++) {
        int f0 = rnd.nextInt() % numInserts;
        int f1 = 5;
        TupleUtils.createIntegerTuple(tb, tuple, f0, f1);
        if (LOGGER.isLoggable(Level.INFO)) {
            if (i % 1000 == 0) {
                LOGGER.info("Inserting " + i + " : " + f0 + " " + f1);
            }
        }
        try {
            indexAccessor.insert(tuple);
        } catch (HyracksDataException e) {
            if (e.getErrorCode() != ErrorCode.DUPLICATE_KEY) {
                throw e;
            }
        }
    }
    long end = System.currentTimeMillis();
    if (LOGGER.isLoggable(Level.INFO)) {
        LOGGER.info(numInserts + " inserts in " + (end - start) + "ms");
    }
    orderedScan(indexAccessor, fieldSerdes);
    diskOrderScan(indexAccessor, fieldSerdes);
    // Build low key.
    ArrayTupleBuilder lowKeyTb = new ArrayTupleBuilder(keyFieldCount);
    ArrayTupleReference lowKey = new ArrayTupleReference();
    TupleUtils.createIntegerTuple(lowKeyTb, lowKey, -1000);
    // Build high key.
    ArrayTupleBuilder highKeyTb = new ArrayTupleBuilder(keyFieldCount);
    ArrayTupleReference highKey = new ArrayTupleReference();
    TupleUtils.createIntegerTuple(highKeyTb, highKey, 1000);
    rangeSearch(cmpFactories, indexAccessor, fieldSerdes, lowKey, highKey, null, null);
    treeIndex.validate();
    treeIndex.deactivate();
    treeIndex.destroy();
}
Also used : ITypeTraits(org.apache.hyracks.api.dataflow.value.ITypeTraits) ArrayTupleReference(org.apache.hyracks.dataflow.common.comm.io.ArrayTupleReference) IBinaryComparatorFactory(org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory) ArrayTupleBuilder(org.apache.hyracks.dataflow.common.comm.io.ArrayTupleBuilder) ISerializerDeserializer(org.apache.hyracks.api.dataflow.value.ISerializerDeserializer) IIndexAccessor(org.apache.hyracks.storage.common.IIndexAccessor) HyracksDataException(org.apache.hyracks.api.exceptions.HyracksDataException) ITreeIndex(org.apache.hyracks.storage.am.common.api.ITreeIndex) Test(org.junit.Test)

Example 27 with IBinaryComparatorFactory

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

the class OrderedIndexExamplesTest method varLenKeyValueExample.

/**
     * Variable-Length Example. Create a BTree with one variable-length key
     * field and one variable-length value field. Fill BTree with random values
     * using insertions (not bulk load) Perform ordered scans and range search.
     */
@Test
public void varLenKeyValueExample() throws Exception {
    if (LOGGER.isLoggable(Level.INFO)) {
        LOGGER.info("Variable-Length Key,Value Example");
    }
    // Declare fields.
    int fieldCount = 2;
    ITypeTraits[] typeTraits = new ITypeTraits[fieldCount];
    typeTraits[0] = UTF8StringPointable.TYPE_TRAITS;
    typeTraits[1] = UTF8StringPointable.TYPE_TRAITS;
    // Declare field serdes.
    ISerializerDeserializer[] fieldSerdes = { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() };
    // Declare keys.
    int keyFieldCount = 1;
    IBinaryComparatorFactory[] cmpFactories = new IBinaryComparatorFactory[keyFieldCount];
    cmpFactories[0] = PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY);
    // This is only used for the LSM-BTree.
    int[] bloomFilterKeyFields = new int[keyFieldCount];
    bloomFilterKeyFields[0] = 0;
    ITreeIndex treeIndex = createTreeIndex(typeTraits, cmpFactories, bloomFilterKeyFields, null, null, null, null);
    treeIndex.create();
    treeIndex.activate();
    long start = System.currentTimeMillis();
    if (LOGGER.isLoggable(Level.INFO)) {
        LOGGER.info("Inserting into tree...");
    }
    ArrayTupleBuilder tb = new ArrayTupleBuilder(fieldCount);
    ArrayTupleReference tuple = new ArrayTupleReference();
    IIndexAccessor indexAccessor = treeIndex.createAccessor(TestOperationCallback.INSTANCE, TestOperationCallback.INSTANCE);
    // Max string length to be generated.
    int maxLength = 10;
    int numInserts = 10000;
    for (int i = 0; i < 10000; i++) {
        String f0 = randomString(Math.abs(rnd.nextInt()) % maxLength + 1, rnd);
        String f1 = randomString(Math.abs(rnd.nextInt()) % maxLength + 1, rnd);
        TupleUtils.createTuple(tb, tuple, fieldSerdes, f0, f1);
        if (LOGGER.isLoggable(Level.INFO)) {
            if (i % 1000 == 0) {
                LOGGER.info("Inserting[" + i + "] " + f0 + " " + f1);
            }
        }
        try {
            indexAccessor.insert(tuple);
        } catch (HyracksDataException e) {
            if (e.getErrorCode() != ErrorCode.DUPLICATE_KEY) {
                throw e;
            }
        }
    }
    long end = System.currentTimeMillis();
    if (LOGGER.isLoggable(Level.INFO)) {
        LOGGER.info(numInserts + " inserts in " + (end - start) + "ms");
    }
    orderedScan(indexAccessor, fieldSerdes);
    diskOrderScan(indexAccessor, fieldSerdes);
    // Build low key.
    ArrayTupleBuilder lowKeyTb = new ArrayTupleBuilder(1);
    ArrayTupleReference lowKey = new ArrayTupleReference();
    TupleUtils.createTuple(lowKeyTb, lowKey, fieldSerdes, "cbf");
    // Build high key.
    ArrayTupleBuilder highKeyTb = new ArrayTupleBuilder(1);
    ArrayTupleReference highKey = new ArrayTupleReference();
    TupleUtils.createTuple(highKeyTb, highKey, fieldSerdes, "cc7");
    rangeSearch(cmpFactories, indexAccessor, fieldSerdes, lowKey, highKey, null, null);
    treeIndex.validate();
    treeIndex.deactivate();
    treeIndex.destroy();
}
Also used : ITypeTraits(org.apache.hyracks.api.dataflow.value.ITypeTraits) ArrayTupleReference(org.apache.hyracks.dataflow.common.comm.io.ArrayTupleReference) IBinaryComparatorFactory(org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory) ArrayTupleBuilder(org.apache.hyracks.dataflow.common.comm.io.ArrayTupleBuilder) UTF8StringSerializerDeserializer(org.apache.hyracks.dataflow.common.data.marshalling.UTF8StringSerializerDeserializer) ISerializerDeserializer(org.apache.hyracks.api.dataflow.value.ISerializerDeserializer) IIndexAccessor(org.apache.hyracks.storage.common.IIndexAccessor) HyracksDataException(org.apache.hyracks.api.exceptions.HyracksDataException) ITreeIndex(org.apache.hyracks.storage.am.common.api.ITreeIndex) Test(org.junit.Test)

Example 28 with IBinaryComparatorFactory

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

the class RangePartitionExchangePOperator method createConnectorDescriptor.

@Override
public Pair<IConnectorDescriptor, TargetConstraint> createConnectorDescriptor(IConnectorDescriptorRegistry spec, ILogicalOperator op, IOperatorSchema opSchema, JobGenContext context) throws AlgebricksException {
    int n = partitioningFields.size();
    int[] sortFields = new int[n];
    IBinaryComparatorFactory[] comps = new IBinaryComparatorFactory[n];
    INormalizedKeyComputerFactoryProvider nkcfProvider = context.getNormalizedKeyComputerFactoryProvider();
    INormalizedKeyComputerFactory nkcf = null;
    IVariableTypeEnvironment env = context.getTypeEnvironment(op);
    int i = 0;
    for (OrderColumn oc : partitioningFields) {
        LogicalVariable var = oc.getColumn();
        sortFields[i] = opSchema.findVariable(var);
        Object type = env.getVarType(var);
        OrderKind order = oc.getOrder();
        if (i == 0 && nkcfProvider != null && type != null) {
            nkcf = nkcfProvider.getNormalizedKeyComputerFactory(type, order == OrderKind.ASC);
        }
        IBinaryComparatorFactoryProvider bcfp = context.getBinaryComparatorFactoryProvider();
        comps[i] = bcfp.getBinaryComparatorFactory(type, oc.getOrder() == OrderKind.ASC);
        i++;
    }
    ITuplePartitionComputerFactory tpcf = new FieldRangePartitionComputerFactory(sortFields, comps, rangeMap);
    IConnectorDescriptor conn = new MToNPartitioningConnectorDescriptor(spec, tpcf);
    return new Pair<IConnectorDescriptor, TargetConstraint>(conn, null);
}
Also used : LogicalVariable(org.apache.hyracks.algebricks.core.algebra.base.LogicalVariable) ITuplePartitionComputerFactory(org.apache.hyracks.api.dataflow.value.ITuplePartitionComputerFactory) IConnectorDescriptor(org.apache.hyracks.api.dataflow.IConnectorDescriptor) FieldRangePartitionComputerFactory(org.apache.hyracks.dataflow.common.data.partition.range.FieldRangePartitionComputerFactory) OrderColumn(org.apache.hyracks.algebricks.core.algebra.properties.OrderColumn) IBinaryComparatorFactory(org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory) MToNPartitioningConnectorDescriptor(org.apache.hyracks.dataflow.std.connectors.MToNPartitioningConnectorDescriptor) IBinaryComparatorFactoryProvider(org.apache.hyracks.algebricks.data.IBinaryComparatorFactoryProvider) TargetConstraint(org.apache.hyracks.algebricks.core.algebra.base.IHyracksJobBuilder.TargetConstraint) INormalizedKeyComputerFactory(org.apache.hyracks.api.dataflow.value.INormalizedKeyComputerFactory) OrderKind(org.apache.hyracks.algebricks.core.algebra.operators.logical.OrderOperator.IOrder.OrderKind) INormalizedKeyComputerFactoryProvider(org.apache.hyracks.algebricks.data.INormalizedKeyComputerFactoryProvider) IVariableTypeEnvironment(org.apache.hyracks.algebricks.core.algebra.expressions.IVariableTypeEnvironment) Pair(org.apache.hyracks.algebricks.common.utils.Pair)

Example 29 with IBinaryComparatorFactory

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

the class RangePartitionMergeExchangePOperator method createConnectorDescriptor.

@Override
public Pair<IConnectorDescriptor, TargetConstraint> createConnectorDescriptor(IConnectorDescriptorRegistry spec, ILogicalOperator op, IOperatorSchema opSchema, JobGenContext context) throws AlgebricksException {
    int n = partitioningFields.size();
    int[] sortFields = new int[n];
    IBinaryComparatorFactory[] comps = new IBinaryComparatorFactory[n];
    INormalizedKeyComputerFactoryProvider nkcfProvider = context.getNormalizedKeyComputerFactoryProvider();
    INormalizedKeyComputerFactory nkcf = null;
    IVariableTypeEnvironment env = context.getTypeEnvironment(op);
    int i = 0;
    for (OrderColumn oc : partitioningFields) {
        LogicalVariable var = oc.getColumn();
        sortFields[i] = opSchema.findVariable(var);
        Object type = env.getVarType(var);
        OrderKind order = oc.getOrder();
        if (i == 0 && nkcfProvider != null && type != null) {
            nkcf = nkcfProvider.getNormalizedKeyComputerFactory(type, order == OrderKind.ASC);
        }
        IBinaryComparatorFactoryProvider bcfp = context.getBinaryComparatorFactoryProvider();
        comps[i] = bcfp.getBinaryComparatorFactory(type, oc.getOrder() == OrderKind.ASC);
        i++;
    }
    ITuplePartitionComputerFactory tpcf = new FieldRangePartitionComputerFactory(sortFields, comps, rangeMap);
    IConnectorDescriptor conn = new MToNPartitioningMergingConnectorDescriptor(spec, tpcf, sortFields, comps, nkcf);
    return new Pair<IConnectorDescriptor, TargetConstraint>(conn, null);
}
Also used : LogicalVariable(org.apache.hyracks.algebricks.core.algebra.base.LogicalVariable) ITuplePartitionComputerFactory(org.apache.hyracks.api.dataflow.value.ITuplePartitionComputerFactory) IConnectorDescriptor(org.apache.hyracks.api.dataflow.IConnectorDescriptor) FieldRangePartitionComputerFactory(org.apache.hyracks.dataflow.common.data.partition.range.FieldRangePartitionComputerFactory) MToNPartitioningMergingConnectorDescriptor(org.apache.hyracks.dataflow.std.connectors.MToNPartitioningMergingConnectorDescriptor) OrderColumn(org.apache.hyracks.algebricks.core.algebra.properties.OrderColumn) IBinaryComparatorFactory(org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory) IBinaryComparatorFactoryProvider(org.apache.hyracks.algebricks.data.IBinaryComparatorFactoryProvider) TargetConstraint(org.apache.hyracks.algebricks.core.algebra.base.IHyracksJobBuilder.TargetConstraint) INormalizedKeyComputerFactory(org.apache.hyracks.api.dataflow.value.INormalizedKeyComputerFactory) OrderKind(org.apache.hyracks.algebricks.core.algebra.operators.logical.OrderOperator.IOrder.OrderKind) INormalizedKeyComputerFactoryProvider(org.apache.hyracks.algebricks.data.INormalizedKeyComputerFactoryProvider) IVariableTypeEnvironment(org.apache.hyracks.algebricks.core.algebra.expressions.IVariableTypeEnvironment) Pair(org.apache.hyracks.algebricks.common.utils.Pair)

Example 30 with IBinaryComparatorFactory

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

the class StableSortPOperator method contributeRuntimeOperator.

@Override
public void contributeRuntimeOperator(IHyracksJobBuilder builder, JobGenContext context, ILogicalOperator op, IOperatorSchema opSchema, IOperatorSchema[] inputSchemas, IOperatorSchema outerPlanSchema) throws AlgebricksException {
    IOperatorDescriptorRegistry spec = builder.getJobSpec();
    RecordDescriptor recDescriptor = JobGenHelper.mkRecordDescriptor(context.getTypeEnvironment(op), opSchema, context);
    int n = sortColumns.length;
    int[] sortFields = new int[n];
    IBinaryComparatorFactory[] comps = new IBinaryComparatorFactory[n];
    INormalizedKeyComputerFactoryProvider nkcfProvider = context.getNormalizedKeyComputerFactoryProvider();
    INormalizedKeyComputerFactory nkcf = null;
    IVariableTypeEnvironment env = context.getTypeEnvironment(op);
    int i = 0;
    for (OrderColumn oc : sortColumns) {
        LogicalVariable var = oc.getColumn();
        sortFields[i] = opSchema.findVariable(var);
        Object type = env.getVarType(var);
        OrderKind order = oc.getOrder();
        if (i == 0 && nkcfProvider != null && type != null) {
            nkcf = nkcfProvider.getNormalizedKeyComputerFactory(type, order == OrderKind.ASC);
        }
        IBinaryComparatorFactoryProvider bcfp = context.getBinaryComparatorFactoryProvider();
        comps[i] = bcfp.getBinaryComparatorFactory(type, oc.getOrder() == OrderKind.ASC);
        i++;
    }
    // topK == -1 means that a topK value is not provided.
    if (topK == -1) {
        ExternalSortOperatorDescriptor sortOpDesc = new ExternalSortOperatorDescriptor(spec, maxNumberOfFrames, sortFields, nkcf, comps, recDescriptor);
        contributeOpDesc(builder, (AbstractLogicalOperator) op, sortOpDesc);
        ILogicalOperator src = op.getInputs().get(0).getValue();
        builder.contributeGraphEdge(src, 0, op, 0);
    } else {
        // Since topK value is provided, topK optimization is possible.
        // We call topKSorter instead of calling ExternalSortOperator.
        TopKSorterOperatorDescriptor sortOpDesc = new TopKSorterOperatorDescriptor(spec, maxNumberOfFrames, topK, sortFields, nkcf, comps, recDescriptor);
        contributeOpDesc(builder, (AbstractLogicalOperator) op, sortOpDesc);
        ILogicalOperator src = op.getInputs().get(0).getValue();
        builder.contributeGraphEdge(src, 0, op, 0);
    }
}
Also used : LogicalVariable(org.apache.hyracks.algebricks.core.algebra.base.LogicalVariable) RecordDescriptor(org.apache.hyracks.api.dataflow.value.RecordDescriptor) OrderColumn(org.apache.hyracks.algebricks.core.algebra.properties.OrderColumn) ILogicalOperator(org.apache.hyracks.algebricks.core.algebra.base.ILogicalOperator) IBinaryComparatorFactory(org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory) IOperatorDescriptorRegistry(org.apache.hyracks.api.job.IOperatorDescriptorRegistry) IBinaryComparatorFactoryProvider(org.apache.hyracks.algebricks.data.IBinaryComparatorFactoryProvider) TopKSorterOperatorDescriptor(org.apache.hyracks.dataflow.std.sort.TopKSorterOperatorDescriptor) INormalizedKeyComputerFactory(org.apache.hyracks.api.dataflow.value.INormalizedKeyComputerFactory) OrderKind(org.apache.hyracks.algebricks.core.algebra.operators.logical.OrderOperator.IOrder.OrderKind) INormalizedKeyComputerFactoryProvider(org.apache.hyracks.algebricks.data.INormalizedKeyComputerFactoryProvider) ExternalSortOperatorDescriptor(org.apache.hyracks.dataflow.std.sort.ExternalSortOperatorDescriptor) IVariableTypeEnvironment(org.apache.hyracks.algebricks.core.algebra.expressions.IVariableTypeEnvironment)

Aggregations

IBinaryComparatorFactory (org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory)86 ITypeTraits (org.apache.hyracks.api.dataflow.value.ITypeTraits)45 RecordDescriptor (org.apache.hyracks.api.dataflow.value.RecordDescriptor)25 ArrayTupleBuilder (org.apache.hyracks.dataflow.common.comm.io.ArrayTupleBuilder)25 Test (org.junit.Test)25 ISerializerDeserializer (org.apache.hyracks.api.dataflow.value.ISerializerDeserializer)24 ArrayTupleReference (org.apache.hyracks.dataflow.common.comm.io.ArrayTupleReference)22 HyracksDataException (org.apache.hyracks.api.exceptions.HyracksDataException)19 UTF8StringSerializerDeserializer (org.apache.hyracks.dataflow.common.data.marshalling.UTF8StringSerializerDeserializer)19 ITreeIndex (org.apache.hyracks.storage.am.common.api.ITreeIndex)17 IIndexAccessor (org.apache.hyracks.storage.common.IIndexAccessor)16 JobSpecification (org.apache.hyracks.api.job.JobSpecification)15 OneToOneConnectorDescriptor (org.apache.hyracks.dataflow.std.connectors.OneToOneConnectorDescriptor)15 IBinaryComparatorFactoryProvider (org.apache.hyracks.algebricks.data.IBinaryComparatorFactoryProvider)14 FieldHashPartitionComputerFactory (org.apache.hyracks.dataflow.common.data.partition.FieldHashPartitionComputerFactory)14 IFileSplitProvider (org.apache.hyracks.dataflow.std.file.IFileSplitProvider)13 LogicalVariable (org.apache.hyracks.algebricks.core.algebra.base.LogicalVariable)11 IVariableTypeEnvironment (org.apache.hyracks.algebricks.core.algebra.expressions.IVariableTypeEnvironment)11 MToNPartitioningMergingConnectorDescriptor (org.apache.hyracks.dataflow.std.connectors.MToNPartitioningMergingConnectorDescriptor)10 IPrimitiveValueProviderFactory (org.apache.hyracks.storage.am.common.api.IPrimitiveValueProviderFactory)10