use of org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory in project asterixdb by apache.
the class LSMBTreeExamplesTest method additionalFilteringingExample.
/**
* Test the LSM component filters.
*/
@Test
public void additionalFilteringingExample() throws Exception {
if (LOGGER.isLoggable(Level.INFO)) {
LOGGER.info("Testing LSMBTree component filters.");
}
// 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;
ITypeTraits[] filterTypeTraits = { IntegerPointable.TYPE_TRAITS };
IBinaryComparatorFactory[] filterCmpFactories = { PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY) };
int[] filterFields = { 1 };
int[] btreeFields = { 1 };
ITreeIndex treeIndex = createTreeIndex(typeTraits, cmpFactories, bloomFilterKeyFields, filterTypeTraits, filterCmpFactories, btreeFields, filterFields);
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 = i;
TupleUtils.createIntegerTuple(tb, tuple, f0, f1);
if (LOGGER.isLoggable(Level.INFO)) {
if (i % 1000 == 0) {
LOGGER.info("Inserting " + i + " : " + f0 + " " + f1);
}
}
indexAccessor.insert(tuple);
}
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);
// Build min filter key.
ArrayTupleBuilder minFilterTb = new ArrayTupleBuilder(filterFields.length);
ArrayTupleReference minTuple = new ArrayTupleReference();
TupleUtils.createIntegerTuple(minFilterTb, minTuple, 400);
// Build max filter key.
ArrayTupleBuilder maxFilterTb = new ArrayTupleBuilder(filterFields.length);
ArrayTupleReference maxTuple = new ArrayTupleReference();
TupleUtils.createIntegerTuple(maxFilterTb, maxTuple, 500);
rangeSearch(cmpFactories, indexAccessor, fieldSerdes, lowKey, highKey, minTuple, maxTuple);
treeIndex.validate();
treeIndex.deactivate();
treeIndex.destroy();
}
use of org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory in project asterixdb by apache.
the class BTreeSearchCursorTest method nonUniqueIndexTest.
@Test
public void nonUniqueIndexTest() throws Exception {
if (LOGGER.isLoggable(Level.INFO)) {
LOGGER.info("TESTING RANGE SEARCH CURSOR ON NONUNIQUE INDEX");
}
IBufferCache bufferCache = harness.getBufferCache();
// declare keys
int keyFieldCount = 2;
IBinaryComparatorFactory[] cmpFactories = new IBinaryComparatorFactory[keyFieldCount];
cmpFactories[0] = PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY);
cmpFactories[1] = PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY);
ITreeIndexFrameFactory leafFrameFactory = new BTreeNSMLeafFrameFactory(tupleWriterFactory);
ITreeIndexFrameFactory interiorFrameFactory = new BTreeNSMInteriorFrameFactory(tupleWriterFactory);
IBTreeLeafFrame leafFrame = (IBTreeLeafFrame) leafFrameFactory.createFrame();
IBTreeInteriorFrame interiorFrame = (IBTreeInteriorFrame) interiorFrameFactory.createFrame();
IMetadataPageManager freePageManager = new LinkedMetaDataPageManager(bufferCache, metaFrameFactory);
BTree btree = new BTree(bufferCache, harness.getFileMapProvider(), freePageManager, interiorFrameFactory, leafFrameFactory, cmpFactories, fieldCount, harness.getFileReference());
btree.create();
btree.activate();
ArrayTupleBuilder tupleBuilder = new ArrayTupleBuilder(fieldCount);
ArrayTupleReference tuple = new ArrayTupleReference();
ITreeIndexAccessor indexAccessor = btree.createAccessor(TestOperationCallback.INSTANCE, TestOperationCallback.INSTANCE);
// generate keys
int numKeys = 50;
int maxKey = 10;
ArrayList<Integer> keys = new ArrayList<>();
for (int i = 0; i < numKeys; i++) {
int k = rnd.nextInt() % maxKey;
keys.add(k);
}
Collections.sort(keys);
// insert keys into btree
for (int i = 0; i < keys.size(); i++) {
TupleUtils.createIntegerTuple(tupleBuilder, tuple, keys.get(i), i);
tuple.reset(tupleBuilder.getFieldEndOffsets(), tupleBuilder.getByteArray());
try {
indexAccessor.insert(tuple);
} catch (Exception e) {
e.printStackTrace();
}
}
int minSearchKey = -100;
int maxSearchKey = 100;
// forward searches
Assert.assertTrue(performSearches(keys, btree, leafFrame, interiorFrame, minSearchKey, maxSearchKey, true, true, false));
Assert.assertTrue(performSearches(keys, btree, leafFrame, interiorFrame, minSearchKey, maxSearchKey, false, true, false));
Assert.assertTrue(performSearches(keys, btree, leafFrame, interiorFrame, minSearchKey, maxSearchKey, true, false, false));
Assert.assertTrue(performSearches(keys, btree, leafFrame, interiorFrame, minSearchKey, maxSearchKey, true, true, false));
btree.deactivate();
btree.destroy();
}
use of org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory in project asterixdb by apache.
the class BTreeSearchCursorTest method uniqueIndexTest.
@Test
public void uniqueIndexTest() throws Exception {
if (LOGGER.isLoggable(Level.INFO)) {
LOGGER.info("TESTING RANGE SEARCH CURSOR ON UNIQUE INDEX");
}
IBufferCache bufferCache = harness.getBufferCache();
// declare keys
int keyFieldCount = 1;
IBinaryComparatorFactory[] cmpFactories = new IBinaryComparatorFactory[keyFieldCount];
cmpFactories[0] = PointableBinaryComparatorFactory.of(IntegerPointable.FACTORY);
ITreeIndexFrameFactory leafFrameFactory = new BTreeNSMLeafFrameFactory(tupleWriterFactory);
ITreeIndexFrameFactory interiorFrameFactory = new BTreeNSMInteriorFrameFactory(tupleWriterFactory);
IBTreeLeafFrame leafFrame = (IBTreeLeafFrame) leafFrameFactory.createFrame();
IBTreeInteriorFrame interiorFrame = (IBTreeInteriorFrame) interiorFrameFactory.createFrame();
IMetadataPageManager freePageManager = new LinkedMetaDataPageManager(bufferCache, metaFrameFactory);
BTree btree = new BTree(bufferCache, harness.getFileMapProvider(), freePageManager, interiorFrameFactory, leafFrameFactory, cmpFactories, fieldCount, harness.getFileReference());
btree.create();
btree.activate();
ArrayTupleBuilder tupleBuilder = new ArrayTupleBuilder(fieldCount);
ArrayTupleReference tuple = new ArrayTupleReference();
ITreeIndexAccessor indexAccessor = btree.createAccessor(TestOperationCallback.INSTANCE, TestOperationCallback.INSTANCE);
// generate keys
int numKeys = 50;
int maxKey = 1000;
TreeSet<Integer> uniqueKeys = new TreeSet<>();
ArrayList<Integer> keys = new ArrayList<>();
while (uniqueKeys.size() < numKeys) {
int key = rnd.nextInt() % maxKey;
uniqueKeys.add(key);
}
for (Integer i : uniqueKeys) {
keys.add(i);
}
// insert keys into btree
for (int i = 0; i < keys.size(); i++) {
TupleUtils.createIntegerTuple(tupleBuilder, tuple, keys.get(i), i);
tuple.reset(tupleBuilder.getFieldEndOffsets(), tupleBuilder.getByteArray());
try {
indexAccessor.insert(tuple);
} catch (Exception e) {
e.printStackTrace();
}
}
int minSearchKey = -100;
int maxSearchKey = 100;
// forward searches
Assert.assertTrue(performSearches(keys, btree, leafFrame, interiorFrame, minSearchKey, maxSearchKey, true, true, false));
Assert.assertTrue(performSearches(keys, btree, leafFrame, interiorFrame, minSearchKey, maxSearchKey, false, true, false));
Assert.assertTrue(performSearches(keys, btree, leafFrame, interiorFrame, minSearchKey, maxSearchKey, true, false, false));
Assert.assertTrue(performSearches(keys, btree, leafFrame, interiorFrame, minSearchKey, maxSearchKey, true, true, false));
btree.deactivate();
btree.destroy();
}
use of org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory in project asterixdb by apache.
the class RTreeResourceFactoryProvider method getCmpFactories.
private static IBinaryComparatorFactory[] getCmpFactories(MetadataProvider metadataProvider, Index index, ARecordType recordType, ARecordType metaType) throws AlgebricksException {
IBinaryComparatorFactoryProvider cmpFactoryProvider = metadataProvider.getStorageComponentProvider().getComparatorFactoryProvider();
List<List<String>> secondaryKeyFields = index.getKeyFieldNames();
int numSecondaryKeys = secondaryKeyFields.size();
if (numSecondaryKeys != 1) {
throw new AsterixException("Cannot use " + numSecondaryKeys + " fields as a key for the R-tree index. " + "There can be only one field as a key for the R-tree index.");
}
List<Integer> keySourceIndicators = index.getKeyFieldSourceIndicators();
ARecordType sourceType;
if (keySourceIndicators == null || keySourceIndicators.get(0) == 0) {
sourceType = recordType;
} else {
sourceType = metaType;
}
Pair<IAType, Boolean> spatialTypePair = Index.getNonNullableOpenFieldType(index.getKeyFieldTypes().get(0), secondaryKeyFields.get(0), sourceType);
IAType spatialType = spatialTypePair.first;
if (spatialType == null) {
throw new AsterixException("Could not find field " + secondaryKeyFields.get(0) + " in the schema.");
}
IAType nestedKeyType = NonTaggedFormatUtil.getNestedSpatialType(spatialType.getTypeTag());
int numDimensions = NonTaggedFormatUtil.getNumDimensions(spatialType.getTypeTag());
int numNestedSecondaryKeyFields = numDimensions * 2;
IBinaryComparatorFactory[] secondaryComparatorFactories = new IBinaryComparatorFactory[numNestedSecondaryKeyFields];
for (int i = 0; i < numNestedSecondaryKeyFields; i++) {
secondaryComparatorFactories[i] = cmpFactoryProvider.getBinaryComparatorFactory(nestedKeyType, true);
}
return secondaryComparatorFactories;
}
use of org.apache.hyracks.api.dataflow.value.IBinaryComparatorFactory 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());
}
Aggregations