use of org.apache.hyracks.api.dataflow.value.IMissingWriterFactory in project asterixdb by apache.
the class TPCHCustomerOrderHashJoinTest method customerOrderCIDHybridHashLeftOuterJoin.
@Test
public void customerOrderCIDHybridHashLeftOuterJoin() 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);
IMissingWriterFactory[] nonMatchWriterFactories = new IMissingWriterFactory[ordersDesc.getFieldCount()];
for (int j = 0; j < nonMatchWriterFactories.length; j++) {
nonMatchWriterFactories[j] = NoopMissingWriterFactory.INSTANCE;
}
HybridHashJoinOperatorDescriptor join = new HybridHashJoinOperatorDescriptor(spec, 32, 20, 200, 1.2, new int[] { 0 }, new int[] { 1 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, custOrderJoinDesc, null, true, nonMatchWriterFactories);
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, 1);
IConnectorDescriptor custJoinConn = new OneToOneConnectorDescriptor(spec);
spec.connect(custJoinConn, custScanner, 0, join, 0);
IConnectorDescriptor joinPrinterConn = new OneToOneConnectorDescriptor(spec);
spec.connect(joinPrinterConn, join, 0, printer, 0);
spec.addRoot(printer);
runTest(spec);
}
use of org.apache.hyracks.api.dataflow.value.IMissingWriterFactory in project asterixdb by apache.
the class TPCHCustomerOrderHashJoinTest method customerOrderCIDInMemoryHashLeftOuterJoin.
@Test
public void customerOrderCIDInMemoryHashLeftOuterJoin() 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);
IMissingWriterFactory[] nonMatchWriterFactories = new IMissingWriterFactory[ordersDesc.getFieldCount()];
for (int j = 0; j < nonMatchWriterFactories.length; j++) {
nonMatchWriterFactories[j] = NoopMissingWriterFactory.INSTANCE;
}
InMemoryHashJoinOperatorDescriptor join = new InMemoryHashJoinOperatorDescriptor(spec, new int[] { 0 }, new int[] { 1 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, null, custOrderJoinDesc, true, nonMatchWriterFactories, 128, 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, 1);
IConnectorDescriptor custJoinConn = new OneToOneConnectorDescriptor(spec);
spec.connect(custJoinConn, custScanner, 0, join, 0);
IConnectorDescriptor joinPrinterConn = new OneToOneConnectorDescriptor(spec);
spec.connect(joinPrinterConn, join, 0, printer, 0);
spec.addRoot(printer);
runTest(spec);
}
use of org.apache.hyracks.api.dataflow.value.IMissingWriterFactory in project asterixdb by apache.
the class TPCHCustomerOrderNestedLoopJoinTest method customerOrderCIDOuterJoinMulti.
@Test
public void customerOrderCIDOuterJoinMulti() throws Exception {
JobSpecification spec = new JobSpecification();
FileSplit[] custSplits = new FileSplit[] { new ManagedFileSplit(NC1_ID, "data" + File.separator + "tpch0.001" + File.separator + "customer-part1.tbl"), new ManagedFileSplit(NC2_ID, "data" + File.separator + "tpch0.001" + File.separator + "customer-part2.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(NC1_ID, "data" + File.separator + "tpch0.001" + File.separator + "orders-part1.tbl"), new ManagedFileSplit(NC2_ID, "data" + File.separator + "tpch0.001" + File.separator + "orders-part2.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, NC1_ID, 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, NC2_ID);
IMissingWriterFactory[] nonMatchWriterFactories = new IMissingWriterFactory[ordersDesc.getFieldCount()];
for (int j = 0; j < nonMatchWriterFactories.length; j++) {
nonMatchWriterFactories[j] = NoopMissingWriterFactory.INSTANCE;
}
NestedLoopJoinOperatorDescriptor join = new NestedLoopJoinOperatorDescriptor(spec, new JoinComparatorFactory(PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), 1, 0), custOrderJoinDesc, 5, true, nonMatchWriterFactories);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, join, NC1_ID, NC2_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 OneToOneConnectorDescriptor(spec);
spec.connect(ordJoinConn, ordScanner, 0, join, 0);
IConnectorDescriptor custJoinConn = new MToNBroadcastConnectorDescriptor(spec);
spec.connect(custJoinConn, custScanner, 0, join, 1);
IConnectorDescriptor joinPrinterConn = new MToNBroadcastConnectorDescriptor(spec);
spec.connect(joinPrinterConn, join, 0, printer, 0);
spec.addRoot(printer);
runTest(spec);
}
use of org.apache.hyracks.api.dataflow.value.IMissingWriterFactory in project asterixdb by apache.
the class JoinMultiComparator method generateHashJoinRuntime.
private IOperatorDescriptor generateHashJoinRuntime(JobGenContext context, IOperatorSchema[] inputSchemas, int[] keysLeft, int[] keysRight, IBinaryHashFunctionFactory[] hashFunFactories, IBinaryComparatorFactory[] comparatorFactories, IPredicateEvaluatorFactory predEvaluatorFactory, RecordDescriptor recDescriptor, IOperatorDescriptorRegistry spec) throws AlgebricksException {
IOperatorDescriptor opDesc;
try {
switch(kind) {
case INNER:
opDesc = new HybridHashJoinOperatorDescriptor(spec, getMemSizeInFrames(), maxInputBuildSizeInFrames, aveRecordsPerFrame, getFudgeFactor(), keysLeft, keysRight, hashFunFactories, comparatorFactories, recDescriptor, predEvaluatorFactory, false, null);
break;
case LEFT_OUTER:
IMissingWriterFactory[] nonMatchWriterFactories = new IMissingWriterFactory[inputSchemas[1].getSize()];
for (int j = 0; j < nonMatchWriterFactories.length; j++) {
nonMatchWriterFactories[j] = context.getMissingWriterFactory();
}
opDesc = new HybridHashJoinOperatorDescriptor(spec, getMemSizeInFrames(), maxInputBuildSizeInFrames, aveRecordsPerFrame, getFudgeFactor(), keysLeft, keysRight, hashFunFactories, comparatorFactories, recDescriptor, predEvaluatorFactory, true, nonMatchWriterFactories);
break;
default:
throw new NotImplementedException();
}
} catch (HyracksDataException e) {
throw new AlgebricksException(e);
}
return opDesc;
}
use of org.apache.hyracks.api.dataflow.value.IMissingWriterFactory in project asterixdb by apache.
the class SubplanPOperator method contributeRuntimeOperator.
@Override
public void contributeRuntimeOperator(IHyracksJobBuilder builder, JobGenContext context, ILogicalOperator op, IOperatorSchema opSchema, IOperatorSchema[] inputSchemas, IOperatorSchema outerPlanSchema) throws AlgebricksException {
SubplanOperator subplan = (SubplanOperator) op;
if (subplan.getNestedPlans().size() != 1) {
throw new NotImplementedException("Subplan currently works only for one nested plan with one root.");
}
AlgebricksPipeline[] subplans = compileSubplans(inputSchemas[0], subplan, opSchema, context);
assert subplans.length == 1;
AlgebricksPipeline np = subplans[0];
RecordDescriptor inputRecordDesc = JobGenHelper.mkRecordDescriptor(context.getTypeEnvironment(op.getInputs().get(0).getValue()), inputSchemas[0], context);
IMissingWriterFactory[] missingWriterFactories = new IMissingWriterFactory[np.getOutputWidth()];
for (int i = 0; i < missingWriterFactories.length; i++) {
missingWriterFactories[i] = context.getMissingWriterFactory();
}
SubplanRuntimeFactory runtime = new SubplanRuntimeFactory(np, missingWriterFactories, inputRecordDesc, null);
RecordDescriptor recDesc = JobGenHelper.mkRecordDescriptor(context.getTypeEnvironment(op), opSchema, context);
builder.contributeMicroOperator(subplan, runtime, recDesc);
ILogicalOperator src = op.getInputs().get(0).getValue();
builder.contributeGraphEdge(src, 0, op, 0);
}
Aggregations