use of org.apache.hyracks.dataflow.common.data.partition.FieldHashPartitionComputerFactory in project asterixdb by apache.
the class AggregationTest method singleKeySumPreClusterGroupTest.
@Test
public void singleKeySumPreClusterGroupTest() throws Exception {
JobSpecification spec = new JobSpecification();
FileScanOperatorDescriptor csvScanner = new FileScanOperatorDescriptor(spec, splitProvider, tupleParserFactory, desc);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, csvScanner, NC2_ID);
RecordDescriptor outputRec = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE, IntegerSerializerDeserializer.INSTANCE, FloatSerializerDeserializer.INSTANCE });
int[] keyFields = new int[] { 0 };
PreclusteredGroupOperatorDescriptor grouper = new PreclusteredGroupOperatorDescriptor(spec, keyFields, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(1, true), new IntSumFieldAggregatorFactory(3, true), new FloatSumFieldAggregatorFactory(5, true) }), outputRec);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, grouper, NC2_ID, NC1_ID);
IConnectorDescriptor conn1 = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(keyFields, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
spec.connect(conn1, csvScanner, 0, grouper, 0);
AbstractSingleActivityOperatorDescriptor printer = getPrinter(spec, "singleKeySumInmemGroupTest");
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC2_ID, NC1_ID);
IConnectorDescriptor conn2 = new OneToOneConnectorDescriptor(spec);
spec.connect(conn2, grouper, 0, printer, 0);
spec.addRoot(printer);
runTest(spec);
}
use of org.apache.hyracks.dataflow.common.data.partition.FieldHashPartitionComputerFactory in project asterixdb by apache.
the class SleepOperatorDescriptor method jobWithSleepOp.
private JobSpecification jobWithSleepOp() {
JobSpecification spec = new JobSpecification();
FileSplit[] ordersSplits = new FileSplit[] { new ManagedFileSplit(ASTERIX_IDS[0], "data" + File.separator + "tpch0.001" + File.separator + "orders-part1.tbl") };
IFileSplitProvider ordersSplitsProvider = new ConstantFileSplitProvider(ordersSplits);
RecordDescriptor recordDesc = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() });
// File scan operator.
FileScanOperatorDescriptor scanOp = 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 }, '|'), recordDesc);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, scanOp, ASTERIX_IDS[0]);
// Sleep operator.
SleepOperatorDescriptor sleepOp = new SleepOperatorDescriptor(spec);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, sleepOp, ASTERIX_IDS);
// Sink operator.
SinkOperatorDescriptor sinkOp = new SinkOperatorDescriptor(spec, 1);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, sinkOp, ASTERIX_IDS);
// Hash-repartitioning connector.
IConnectorDescriptor conn1 = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(new int[] { 0 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
spec.connect(conn1, scanOp, 0, sleepOp, 0);
// One-to-one connector.
IConnectorDescriptor conn2 = new OneToOneConnectorDescriptor(spec);
spec.connect(conn2, sleepOp, 0, sinkOp, 0);
return spec;
}
use of org.apache.hyracks.dataflow.common.data.partition.FieldHashPartitionComputerFactory in project asterixdb by apache.
the class HeapSortMergeTest method optimizedSortMergeTest01.
@Test
public void optimizedSortMergeTest01() throws Exception {
JobSpecification spec = new JobSpecification();
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 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() });
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 }, '|'), ordersDesc);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, ordScanner, NC1_ID, NC2_ID);
// larger than the total record numbers.
int outputLimit = 5;
TopKSorterOperatorDescriptor sorter = new TopKSorterOperatorDescriptor(spec, 4, outputLimit, new int[] { 1, 0 }, null, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, ordersDesc);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, sorter, NC1_ID, NC2_ID);
ResultSetId rsId = new ResultSetId(1);
spec.addResultSetId(rsId);
FileSplit fs = createFile(nc1);
IFileSplitProvider outputSplitProvider = new ConstantFileSplitProvider(new FileSplit[] { fs });
IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outputSplitProvider, "|");
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
spec.connect(new OneToOneConnectorDescriptor(spec), ordScanner, 0, sorter, 0);
spec.connect(new MToNPartitioningMergingConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(new int[] { 1, 0 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY), PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }), new int[] { 1, 0 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new UTF8StringNormalizedKeyComputerFactory()), sorter, 0, printer, 0);
runTest(spec);
System.out.println("Result write into :" + fs.getPath() + " in node: " + fs.getNodeName());
}
use of org.apache.hyracks.dataflow.common.data.partition.FieldHashPartitionComputerFactory in project asterixdb by apache.
the class HeapSortMergeTest method createSortMergeJobSpec.
public static JobSpecification createSortMergeJobSpec() throws Exception {
JobSpecification spec = new JobSpecification();
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 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() });
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 }, '|'), ordersDesc);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, ordScanner, NC1_ID, NC2_ID);
int outputLimit = 20;
TopKSorterOperatorDescriptor sorter = new TopKSorterOperatorDescriptor(spec, 4, outputLimit, new int[] { 1, 0 }, null, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, ordersDesc);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, sorter, NC1_ID, NC2_ID);
LimitOperatorDescriptor filter = new LimitOperatorDescriptor(spec, ordersDesc, outputLimit);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, filter, 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);
spec.connect(new OneToOneConnectorDescriptor(spec), ordScanner, 0, sorter, 0);
spec.connect(new MToNPartitioningMergingConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(new int[] { 1, 0 }, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY), PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }), new int[] { 1, 0 }, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new UTF8StringNormalizedKeyComputerFactory()), sorter, 0, filter, 0);
spec.connect(new OneToOneConnectorDescriptor(spec), filter, 0, printer, 0);
return spec;
}
use of org.apache.hyracks.dataflow.common.data.partition.FieldHashPartitionComputerFactory in project asterixdb by apache.
the class AggregationTest method multiKeyAvgExtGroupTest.
@Test
public void multiKeyAvgExtGroupTest() throws Exception {
JobSpecification spec = new JobSpecification();
FileScanOperatorDescriptor csvScanner = new FileScanOperatorDescriptor(spec, splitProvider, tupleParserFactory, desc);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, csvScanner, NC2_ID);
RecordDescriptor outputRec = new RecordDescriptor(new ISerializerDeserializer[] { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer(), IntegerSerializerDeserializer.INSTANCE, IntegerSerializerDeserializer.INSTANCE, FloatSerializerDeserializer.INSTANCE });
int[] keyFields = new int[] { 8, 0 };
int frameLimits = 5;
int tableSize = 8;
long fileSize = frameLimits * spec.getFrameSize();
ExternalGroupOperatorDescriptor grouper = new ExternalGroupOperatorDescriptor(spec, tableSize, fileSize, keyFields, frameLimits, new IBinaryComparatorFactory[] { PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY), PointableBinaryComparatorFactory.of(UTF8StringPointable.FACTORY) }, new UTF8StringNormalizedKeyComputerFactory(), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(1, false), new CountFieldAggregatorFactory(false), new AvgFieldGroupAggregatorFactory(1, false) }), new MultiFieldsAggregatorFactory(new IFieldAggregateDescriptorFactory[] { new IntSumFieldAggregatorFactory(2, false), new IntSumFieldAggregatorFactory(3, false), new AvgFieldMergeAggregatorFactory(4, false) }), outputRec, outputRec, new HashSpillableTableFactory(new IBinaryHashFunctionFamily[] { UTF8StringBinaryHashFunctionFamily.INSTANCE, UTF8StringBinaryHashFunctionFamily.INSTANCE }));
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, grouper, NC2_ID, NC1_ID);
IConnectorDescriptor conn1 = new MToNPartitioningConnectorDescriptor(spec, new FieldHashPartitionComputerFactory(keyFields, new IBinaryHashFunctionFactory[] { PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY), PointableBinaryHashFunctionFactory.of(UTF8StringPointable.FACTORY) }));
spec.connect(conn1, csvScanner, 0, grouper, 0);
AbstractSingleActivityOperatorDescriptor printer = getPrinter(spec, "multiKeyAvgExtGroupTest");
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC2_ID, NC1_ID);
IConnectorDescriptor conn2 = new OneToOneConnectorDescriptor(spec);
spec.connect(conn2, grouper, 0, printer, 0);
spec.addRoot(printer);
runTest(spec);
}
Aggregations