use of org.apache.hyracks.api.dataflow.IOperatorDescriptor 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.api.dataflow.IOperatorDescriptor in project asterixdb by apache.
the class BTreeSecondaryIndexSearchOperatorTest method searchSecondaryIndexTest.
@Test
public void searchSecondaryIndexTest() throws Exception {
JobSpecification spec = new JobSpecification();
// build tuple containing search keys (only use the first key as search
// key)
ArrayTupleBuilder tb = new ArrayTupleBuilder(DataSetConstants.secondaryKeyFieldCount);
DataOutput dos = tb.getDataOutput();
tb.reset();
// low key
new UTF8StringSerializerDeserializer().serialize("1998-07-21", dos);
tb.addFieldEndOffset();
// high key
new UTF8StringSerializerDeserializer().serialize("2000-10-18", dos);
tb.addFieldEndOffset();
ISerializerDeserializer[] keyRecDescSers = { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() };
RecordDescriptor keyRecDesc = new RecordDescriptor(keyRecDescSers);
ConstantTupleSourceOperatorDescriptor keyProviderOp = new ConstantTupleSourceOperatorDescriptor(spec, keyRecDesc, tb.getFieldEndOffsets(), tb.getByteArray(), tb.getSize());
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, keyProviderOp, NC1_ID);
int[] secondaryLowKeyFields = { 0 };
int[] secondaryHighKeyFields = { 1 };
// search secondary index
BTreeSearchOperatorDescriptor secondaryBtreeSearchOp = new BTreeSearchOperatorDescriptor(spec, DataSetConstants.secondaryRecDesc, secondaryLowKeyFields, secondaryHighKeyFields, true, true, secondaryHelperFactory, false, false, null, NoOpOperationCallbackFactory.INSTANCE, null, null, false);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, secondaryBtreeSearchOp, NC1_ID);
// second field from the tuples
int[] primaryLowKeyFields = { 1 };
// coming from secondary index
// second field from the tuples
int[] primaryHighKeyFields = { 1 };
// coming from secondary index
// search primary index
BTreeSearchOperatorDescriptor primaryBtreeSearchOp = new BTreeSearchOperatorDescriptor(spec, DataSetConstants.primaryRecDesc, primaryLowKeyFields, primaryHighKeyFields, true, true, primaryHelperFactory, false, false, null, NoOpOperationCallbackFactory.INSTANCE, null, null, false);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, primaryBtreeSearchOp, NC1_ID);
IFileSplitProvider outSplits = new ConstantFileSplitProvider(new FileSplit[] { createFile(nc1) });
IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outSplits, ",");
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
spec.connect(new OneToOneConnectorDescriptor(spec), keyProviderOp, 0, secondaryBtreeSearchOp, 0);
spec.connect(new OneToOneConnectorDescriptor(spec), secondaryBtreeSearchOp, 0, primaryBtreeSearchOp, 0);
spec.connect(new OneToOneConnectorDescriptor(spec), primaryBtreeSearchOp, 0, printer, 0);
spec.addRoot(printer);
runTest(spec);
}
use of org.apache.hyracks.api.dataflow.IOperatorDescriptor in project asterixdb by apache.
the class BTreeSecondaryIndexUpsertOperatorTest method searchUpdatedSecondaryIndexTest.
@Test
public void searchUpdatedSecondaryIndexTest() throws Exception {
JobSpecification spec = new JobSpecification();
// build tuple containing search keys (only use the first key as search
// key)
ArrayTupleBuilder tb = new ArrayTupleBuilder(DataSetConstants.secondaryKeyFieldCount);
DataOutput dos = tb.getDataOutput();
tb.reset();
// low key
new UTF8StringSerializerDeserializer().serialize("1998-07-21", dos);
tb.addFieldEndOffset();
// high key
new UTF8StringSerializerDeserializer().serialize("2000-10-18", dos);
tb.addFieldEndOffset();
ISerializerDeserializer[] keyRecDescSers = { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() };
RecordDescriptor keyRecDesc = new RecordDescriptor(keyRecDescSers);
ConstantTupleSourceOperatorDescriptor keyProviderOp = new ConstantTupleSourceOperatorDescriptor(spec, keyRecDesc, tb.getFieldEndOffsets(), tb.getByteArray(), tb.getSize());
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, keyProviderOp, NC1_ID);
int[] secondaryLowKeyFields = { 0 };
int[] secondaryHighKeyFields = { 1 };
// search secondary index
BTreeSearchOperatorDescriptor secondaryBtreeSearchOp = new BTreeSearchOperatorDescriptor(spec, DataSetConstants.secondaryRecDesc, secondaryLowKeyFields, secondaryHighKeyFields, true, true, secondaryHelperFactory, false, false, null, NoOpOperationCallbackFactory.INSTANCE, null, null, false);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, secondaryBtreeSearchOp, NC1_ID);
// second field from the tuples coming from secondary index
int[] primaryLowKeyFields = { 1 };
// second field from the tuples coming from secondary index
int[] primaryHighKeyFields = { 1 };
// search primary index
BTreeSearchOperatorDescriptor primaryBtreeSearchOp = new BTreeSearchOperatorDescriptor(spec, DataSetConstants.primaryRecDesc, primaryLowKeyFields, primaryHighKeyFields, true, true, primaryHelperFactory, false, false, null, NoOpOperationCallbackFactory.INSTANCE, null, null, false);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, primaryBtreeSearchOp, NC1_ID);
IFileSplitProvider outSplits = new ConstantFileSplitProvider(new FileSplit[] { createFile(nc1) });
IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outSplits, ",");
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
spec.connect(new OneToOneConnectorDescriptor(spec), keyProviderOp, 0, secondaryBtreeSearchOp, 0);
spec.connect(new OneToOneConnectorDescriptor(spec), secondaryBtreeSearchOp, 0, primaryBtreeSearchOp, 0);
spec.connect(new OneToOneConnectorDescriptor(spec), primaryBtreeSearchOp, 0, printer, 0);
spec.addRoot(printer);
runTest(spec);
}
use of org.apache.hyracks.api.dataflow.IOperatorDescriptor in project asterixdb by apache.
the class LSMBTreePrimaryIndexSearchOperatorTest method shouldWriteFilterValueIfAppendFilterIsTrue.
@Test
public void shouldWriteFilterValueIfAppendFilterIsTrue() throws Exception {
JobSpecification spec = new JobSpecification();
// build tuple containing low and high search key
// high key and low key
ArrayTupleBuilder tb = new ArrayTupleBuilder(DataSetConstants.primaryKeyFieldCount * 2);
DataOutput dos = tb.getDataOutput();
tb.reset();
// low key
new UTF8StringSerializerDeserializer().serialize("100", dos);
tb.addFieldEndOffset();
// high key
new UTF8StringSerializerDeserializer().serialize("200", dos);
tb.addFieldEndOffset();
ISerializerDeserializer[] keyRecDescSers = { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() };
RecordDescriptor keyRecDesc = new RecordDescriptor(keyRecDescSers);
ConstantTupleSourceOperatorDescriptor keyProviderOp = new ConstantTupleSourceOperatorDescriptor(spec, keyRecDesc, tb.getFieldEndOffsets(), tb.getByteArray(), tb.getSize());
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, keyProviderOp, NC1_ID);
int[] lowKeyFields = { 0 };
int[] highKeyFields = { 1 };
BTreeSearchOperatorDescriptor primaryBtreeSearchOp = new BTreeSearchOperatorDescriptor(spec, DataSetConstants.primaryAndFilterRecDesc, lowKeyFields, highKeyFields, true, true, primaryHelperFactory, false, false, NoopMissingWriterFactory.INSTANCE, NoOpOperationCallbackFactory.INSTANCE, null, null, true);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, primaryBtreeSearchOp, NC1_ID);
IFileSplitProvider outSplits = new ConstantFileSplitProvider(new FileSplit[] { createFile(nc1) });
IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outSplits, ",");
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
spec.connect(new OneToOneConnectorDescriptor(spec), keyProviderOp, 0, primaryBtreeSearchOp, 0);
spec.connect(new OneToOneConnectorDescriptor(spec), primaryBtreeSearchOp, 0, printer, 0);
spec.addRoot(printer);
runTest(spec);
}
use of org.apache.hyracks.api.dataflow.IOperatorDescriptor in project asterixdb by apache.
the class LSMBTreeSecondaryIndexSearchOperatorTest method shouldCarryFilterValueToPrimaryIndexSearch.
@Test
public void shouldCarryFilterValueToPrimaryIndexSearch() throws Exception {
JobSpecification spec = new JobSpecification();
// build tuple containing search keys (only use the first key as search
// key)
ArrayTupleBuilder tb = new ArrayTupleBuilder(DataSetConstants.secondaryKeyFieldCount);
DataOutput dos = tb.getDataOutput();
tb.reset();
// low key
new UTF8StringSerializerDeserializer().serialize("1998-07-21", dos);
tb.addFieldEndOffset();
// high key
new UTF8StringSerializerDeserializer().serialize("2000-10-18", dos);
tb.addFieldEndOffset();
ISerializerDeserializer[] keyRecDescSers = { new UTF8StringSerializerDeserializer(), new UTF8StringSerializerDeserializer() };
RecordDescriptor keyRecDesc = new RecordDescriptor(keyRecDescSers);
ConstantTupleSourceOperatorDescriptor keyProviderOp = new ConstantTupleSourceOperatorDescriptor(spec, keyRecDesc, tb.getFieldEndOffsets(), tb.getByteArray(), tb.getSize());
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, keyProviderOp, NC1_ID);
int[] secondaryLowKeyFields = { 0 };
int[] secondaryHighKeyFields = { 1 };
// search secondary index
BTreeSearchOperatorDescriptor secondaryBtreeSearchOp = new BTreeSearchOperatorDescriptor(spec, DataSetConstants.secondaryWithFilterRecDesc, secondaryLowKeyFields, secondaryHighKeyFields, true, true, secondaryHelperFactory, false, false, NoopMissingWriterFactory.INSTANCE, NoOpOperationCallbackFactory.INSTANCE, null, null, true);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, secondaryBtreeSearchOp, NC1_ID);
// second field from the tuples
int[] primaryLowKeyFields = { 1 };
// second field from the tuples
int[] primaryHighKeyFields = { 1 };
int[] minFilterFields = { 2 };
int[] maxFilterFields = { 3 };
// search primary index
BTreeSearchOperatorDescriptor primaryBtreeSearchOp = new BTreeSearchOperatorDescriptor(spec, DataSetConstants.primaryRecDesc, primaryLowKeyFields, primaryHighKeyFields, true, true, primaryHelperFactory, false, false, NoopMissingWriterFactory.INSTANCE, NoOpOperationCallbackFactory.INSTANCE, minFilterFields, maxFilterFields, false);
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, primaryBtreeSearchOp, NC1_ID);
IFileSplitProvider outSplits = new ConstantFileSplitProvider(new FileSplit[] { createFile(nc1) });
IOperatorDescriptor printer = new PlainFileWriterOperatorDescriptor(spec, outSplits, ",");
PartitionConstraintHelper.addAbsoluteLocationConstraint(spec, printer, NC1_ID);
spec.connect(new OneToOneConnectorDescriptor(spec), keyProviderOp, 0, secondaryBtreeSearchOp, 0);
spec.connect(new OneToOneConnectorDescriptor(spec), secondaryBtreeSearchOp, 0, primaryBtreeSearchOp, 0);
spec.connect(new OneToOneConnectorDescriptor(spec), primaryBtreeSearchOp, 0, printer, 0);
spec.addRoot(printer);
runTest(spec);
}
Aggregations