org.apache.flink.api.common.operators.base.ReduceOperatorBase.CombineHint Java Examples

The following examples show how to use org.apache.flink.api.common.operators.base.ReduceOperatorBase.CombineHint. You can vote up the ones you like or vote down the ones you don't like, and go to the original project or source file by following the links above each example. You may check out the related API usage on the sidebar.
Example #1
Source File: ReducePerformance.java    From Flink-CEPplus with Apache License 2.0 6 votes vote down vote up 
private static <T, B extends CopyableIterator<T>> void testReducePerformance
	(B iterator, TypeInformation<T> typeInfo, CombineHint hint, int numRecords, boolean print) throws Exception {

	ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
	env.getConfig().enableObjectReuse();

	@SuppressWarnings("unchecked")
	DataSet<T> output =
		env.fromParallelCollection(new SplittableRandomIterator<T, B>(numRecords, iterator), typeInfo)
			.groupBy("0")
			.reduce(new SumReducer()).setCombineHint(hint);

	long start = System.currentTimeMillis();

	System.out.println(output.count());

	long end = System.currentTimeMillis();
	if (print) {
		System.out.println("=== Time for " + iterator.getClass().getSimpleName() + " with hint " + hint.toString() + ": " + (end - start) + "ms ===");
	}
}
Example #2
Source File: Simplify.java    From flink with Apache License 2.0 6 votes vote down vote up 
@Override
public Graph<K, VV, EV> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// Edges
	DataSet<Edge<K, EV>> edges = input
		.getEdges()
		.flatMap(new SymmetrizeAndRemoveSelfLoops<>(clipAndFlip))
			.setParallelism(parallelism)
			.name("Remove self-loops")
		.distinct(0, 1)
			.setCombineHint(CombineHint.NONE)
			.setParallelism(parallelism)
			.name("Remove duplicate edges");

	// Graph
	return Graph.fromDataSet(input.getVertices(), edges, input.getContext());
}
Example #3
Source File: Simplify.java    From flink with Apache License 2.0 6 votes vote down vote up 
@Override
public Graph<K, VV, EV> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// Edges
	DataSet<Edge<K, EV>> edges = input
		.getEdges()
		.filter(new RemoveSelfLoops<>())
			.setParallelism(parallelism)
			.name("Remove self-loops")
		.distinct(0, 1)
			.setCombineHint(CombineHint.NONE)
			.setParallelism(parallelism)
			.name("Remove duplicate edges");

	// Graph
	return Graph.fromDataSet(input.getVertices(), edges, input.getContext());
}
Example #4
Source File: DistinctOperator.java    From flink with Apache License 2.0 6 votes vote down vote up 
private static <IN, K> org.apache.flink.api.common.operators.SingleInputOperator<?, IN, ?> translateSelectorFunctionDistinct(
		SelectorFunctionKeys<IN, ?> rawKeys,
		ReduceFunction<IN> function,
		TypeInformation<IN> outputType,
		String name,
		Operator<IN> input,
		int parallelism,
		CombineHint hint) {
	@SuppressWarnings("unchecked")
	final SelectorFunctionKeys<IN, K> keys = (SelectorFunctionKeys<IN, K>) rawKeys;

	TypeInformation<Tuple2<K, IN>> typeInfoWithKey = KeyFunctions.createTypeWithKey(keys);
	Operator<Tuple2<K, IN>> keyedInput = KeyFunctions.appendKeyExtractor(input, keys);

	PlanUnwrappingReduceOperator<IN, K> reducer =
			new PlanUnwrappingReduceOperator<>(function, keys, name, outputType, typeInfoWithKey);
	reducer.setInput(keyedInput);
	reducer.setCombineHint(hint);
	reducer.setParallelism(parallelism);

	return KeyFunctions.appendKeyRemover(reducer, keys);
}
Example #5
Source File: ReduceOperator.java    From flink with Apache License 2.0 6 votes vote down vote up 
private static <T, K> org.apache.flink.api.common.operators.SingleInputOperator<?, T, ?> translateSelectorFunctionReducer(
	SelectorFunctionKeys<T, ?> rawKeys,
	ReduceFunction<T> function,
	TypeInformation<T> inputType,
	String name,
	Operator<T> input,
	int parallelism,
	CombineHint hint) {
	@SuppressWarnings("unchecked")
	final SelectorFunctionKeys<T, K> keys = (SelectorFunctionKeys<T, K>) rawKeys;

	TypeInformation<Tuple2<K, T>> typeInfoWithKey = KeyFunctions.createTypeWithKey(keys);
	Operator<Tuple2<K, T>> keyedInput = KeyFunctions.appendKeyExtractor(input, keys);

	PlanUnwrappingReduceOperator<T, K> reducer = new PlanUnwrappingReduceOperator<>(function, keys, name, inputType, typeInfoWithKey);
	reducer.setInput(keyedInput);
	reducer.setParallelism(parallelism);
	reducer.setCombineHint(hint);

	return KeyFunctions.appendKeyRemover(reducer, keys);
}
Example #6
Source File: ReducePerformance.java    From flink with Apache License 2.0 6 votes vote down vote up 
private static <T, B extends CopyableIterator<T>> void testReducePerformance
	(B iterator, TypeInformation<T> typeInfo, CombineHint hint, int numRecords, boolean print) throws Exception {

	ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
	env.getConfig().enableObjectReuse();

	@SuppressWarnings("unchecked")
	DataSet<T> output =
		env.fromParallelCollection(new SplittableRandomIterator<T, B>(numRecords, iterator), typeInfo)
			.groupBy("0")
			.reduce(new SumReducer()).setCombineHint(hint);

	long start = System.currentTimeMillis();

	System.out.println(output.count());

	long end = System.currentTimeMillis();
	if (print) {
		System.out.println("=== Time for " + iterator.getClass().getSimpleName() + " with hint " + hint.toString() + ": " + (end - start) + "ms ===");
	}
}
Example #7
Source File: ReducePerformance.java    From flink with Apache License 2.0 6 votes vote down vote up 
private static <T, B extends CopyableIterator<T>> void testReducePerformance
	(B iterator, TypeInformation<T> typeInfo, CombineHint hint, int numRecords, boolean print) throws Exception {

	ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
	env.getConfig().enableObjectReuse();

	@SuppressWarnings("unchecked")
	DataSet<T> output =
		env.fromParallelCollection(new SplittableRandomIterator<T, B>(numRecords, iterator), typeInfo)
			.groupBy("0")
			.reduce(new SumReducer()).setCombineHint(hint);

	long start = System.currentTimeMillis();

	System.out.println(output.count());

	long end = System.currentTimeMillis();
	if (print) {
		System.out.println("=== Time for " + iterator.getClass().getSimpleName() + " with hint " + hint.toString() + ": " + (end - start) + "ms ===");
	}
}
Example #8
Source File: ReduceOperator.java    From flink with Apache License 2.0 6 votes vote down vote up 
private static <T, K> org.apache.flink.api.common.operators.SingleInputOperator<?, T, ?> translateSelectorFunctionReducer(
	SelectorFunctionKeys<T, ?> rawKeys,
	ReduceFunction<T> function,
	TypeInformation<T> inputType,
	String name,
	Operator<T> input,
	int parallelism,
	CombineHint hint) {
	@SuppressWarnings("unchecked")
	final SelectorFunctionKeys<T, K> keys = (SelectorFunctionKeys<T, K>) rawKeys;

	TypeInformation<Tuple2<K, T>> typeInfoWithKey = KeyFunctions.createTypeWithKey(keys);
	Operator<Tuple2<K, T>> keyedInput = KeyFunctions.appendKeyExtractor(input, keys);

	PlanUnwrappingReduceOperator<T, K> reducer = new PlanUnwrappingReduceOperator<>(function, keys, name, inputType, typeInfoWithKey);
	reducer.setInput(keyedInput);
	reducer.setParallelism(parallelism);
	reducer.setCombineHint(hint);

	return KeyFunctions.appendKeyRemover(reducer, keys);
}
Example #9
Source File: DistinctOperator.java    From flink with Apache License 2.0 6 votes vote down vote up 
private static <IN, K> org.apache.flink.api.common.operators.SingleInputOperator<?, IN, ?> translateSelectorFunctionDistinct(
		SelectorFunctionKeys<IN, ?> rawKeys,
		ReduceFunction<IN> function,
		TypeInformation<IN> outputType,
		String name,
		Operator<IN> input,
		int parallelism,
		CombineHint hint) {
	@SuppressWarnings("unchecked")
	final SelectorFunctionKeys<IN, K> keys = (SelectorFunctionKeys<IN, K>) rawKeys;

	TypeInformation<Tuple2<K, IN>> typeInfoWithKey = KeyFunctions.createTypeWithKey(keys);
	Operator<Tuple2<K, IN>> keyedInput = KeyFunctions.appendKeyExtractor(input, keys);

	PlanUnwrappingReduceOperator<IN, K> reducer =
			new PlanUnwrappingReduceOperator<>(function, keys, name, outputType, typeInfoWithKey);
	reducer.setInput(keyedInput);
	reducer.setCombineHint(hint);
	reducer.setParallelism(parallelism);

	return KeyFunctions.appendKeyRemover(reducer, keys);
}
Example #10
Source File: Simplify.java    From flink with Apache License 2.0 6 votes vote down vote up 
@Override
public Graph<K, VV, EV> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// Edges
	DataSet<Edge<K, EV>> edges = input
		.getEdges()
		.filter(new RemoveSelfLoops<>())
			.setParallelism(parallelism)
			.name("Remove self-loops")
		.distinct(0, 1)
			.setCombineHint(CombineHint.NONE)
			.setParallelism(parallelism)
			.name("Remove duplicate edges");

	// Graph
	return Graph.fromDataSet(input.getVertices(), edges, input.getContext());
}
Example #11
Source File: Simplify.java    From flink with Apache License 2.0 6 votes vote down vote up 
@Override
public Graph<K, VV, EV> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// Edges
	DataSet<Edge<K, EV>> edges = input
		.getEdges()
		.flatMap(new SymmetrizeAndRemoveSelfLoops<>(clipAndFlip))
			.setParallelism(parallelism)
			.name("Remove self-loops")
		.distinct(0, 1)
			.setCombineHint(CombineHint.NONE)
			.setParallelism(parallelism)
			.name("Remove duplicate edges");

	// Graph
	return Graph.fromDataSet(input.getVertices(), edges, input.getContext());
}
Example #12
Source File: Simplify.java    From Flink-CEPplus with Apache License 2.0 6 votes vote down vote up 
@Override
public Graph<K, VV, EV> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// Edges
	DataSet<Edge<K, EV>> edges = input
		.getEdges()
		.flatMap(new SymmetrizeAndRemoveSelfLoops<>(clipAndFlip))
			.setParallelism(parallelism)
			.name("Remove self-loops")
		.distinct(0, 1)
			.setCombineHint(CombineHint.NONE)
			.setParallelism(parallelism)
			.name("Remove duplicate edges");

	// Graph
	return Graph.fromDataSet(input.getVertices(), edges, input.getContext());
}
Example #13
Source File: Simplify.java    From Flink-CEPplus with Apache License 2.0 6 votes vote down vote up 
@Override
public Graph<K, VV, EV> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// Edges
	DataSet<Edge<K, EV>> edges = input
		.getEdges()
		.filter(new RemoveSelfLoops<>())
			.setParallelism(parallelism)
			.name("Remove self-loops")
		.distinct(0, 1)
			.setCombineHint(CombineHint.NONE)
			.setParallelism(parallelism)
			.name("Remove duplicate edges");

	// Graph
	return Graph.fromDataSet(input.getVertices(), edges, input.getContext());
}
Example #14
Source File: DistinctOperator.java    From Flink-CEPplus with Apache License 2.0 6 votes vote down vote up 
private static <IN, K> org.apache.flink.api.common.operators.SingleInputOperator<?, IN, ?> translateSelectorFunctionDistinct(
		SelectorFunctionKeys<IN, ?> rawKeys,
		ReduceFunction<IN> function,
		TypeInformation<IN> outputType,
		String name,
		Operator<IN> input,
		int parallelism,
		CombineHint hint) {
	@SuppressWarnings("unchecked")
	final SelectorFunctionKeys<IN, K> keys = (SelectorFunctionKeys<IN, K>) rawKeys;

	TypeInformation<Tuple2<K, IN>> typeInfoWithKey = KeyFunctions.createTypeWithKey(keys);
	Operator<Tuple2<K, IN>> keyedInput = KeyFunctions.appendKeyExtractor(input, keys);

	PlanUnwrappingReduceOperator<IN, K> reducer =
			new PlanUnwrappingReduceOperator<>(function, keys, name, outputType, typeInfoWithKey);
	reducer.setInput(keyedInput);
	reducer.setCombineHint(hint);
	reducer.setParallelism(parallelism);

	return KeyFunctions.appendKeyRemover(reducer, keys);
}
Example #15
Source File: ReduceOperator.java    From Flink-CEPplus with Apache License 2.0 6 votes vote down vote up 
private static <T, K> org.apache.flink.api.common.operators.SingleInputOperator<?, T, ?> translateSelectorFunctionReducer(
	SelectorFunctionKeys<T, ?> rawKeys,
	ReduceFunction<T> function,
	TypeInformation<T> inputType,
	String name,
	Operator<T> input,
	int parallelism,
	CombineHint hint) {
	@SuppressWarnings("unchecked")
	final SelectorFunctionKeys<T, K> keys = (SelectorFunctionKeys<T, K>) rawKeys;

	TypeInformation<Tuple2<K, T>> typeInfoWithKey = KeyFunctions.createTypeWithKey(keys);
	Operator<Tuple2<K, T>> keyedInput = KeyFunctions.appendKeyExtractor(input, keys);

	PlanUnwrappingReduceOperator<T, K> reducer = new PlanUnwrappingReduceOperator<>(function, keys, name, inputType, typeInfoWithKey);
	reducer.setInput(keyedInput);
	reducer.setParallelism(parallelism);
	reducer.setCombineHint(hint);

	return KeyFunctions.appendKeyRemover(reducer, keys);
}
Example #16
Source File: LocalClusteringCoefficient.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public DataSet<Result<K>> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// u, v, w
	DataSet<TriangleListing.Result<K>> triangles = input
		.run(new TriangleListing<K, VV, EV>()
			.setParallelism(parallelism));

	// u, 1
	DataSet<Tuple2<K, LongValue>> triangleVertices = triangles
		.flatMap(new SplitTriangles<>())
			.name("Split triangle vertices");

	// u, triangle count
	DataSet<Tuple2<K, LongValue>> vertexTriangleCount = triangleVertices
		.groupBy(0)
		.reduce(new CountTriangles<>())
		.setCombineHint(CombineHint.HASH)
			.name("Count triangles")
			.setParallelism(parallelism);

	// u, deg(u)
	DataSet<Vertex<K, LongValue>> vertexDegree = input
		.run(new VertexDegree<K, VV, EV>()
			.setIncludeZeroDegreeVertices(includeZeroDegreeVertices.get())
			.setParallelism(parallelism));

	// u, deg(u), triangle count
	return vertexDegree
		.leftOuterJoin(vertexTriangleCount)
		.where(0)
		.equalTo(0)
		.with(new JoinVertexDegreeWithTriangleCount<>())
			.setParallelism(parallelism)
			.name("Clustering coefficient");
}
Example #17
Source File: VertexDegree.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public DataSet<Vertex<K, LongValue>> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	MapFunction<Edge<K, EV>, Vertex<K, LongValue>> mapEdgeToId = reduceOnTargetId.get() ?
		new MapEdgeToTargetId<>() : new MapEdgeToSourceId<>();

	// v
	DataSet<Vertex<K, LongValue>> vertexIds = input
		.getEdges()
		.map(mapEdgeToId)
			.setParallelism(parallelism)
			.name("Edge to vertex ID");

	// v, deg(v)
	DataSet<Vertex<K, LongValue>> degree = vertexIds
		.groupBy(0)
		.reduce(new DegreeCount<>())
		.setCombineHint(CombineHint.HASH)
			.setParallelism(parallelism)
			.name("Degree count");

	if (includeZeroDegreeVertices.get()) {
		degree = input
			.getVertices()
			.leftOuterJoin(degree)
			.where(0)
			.equalTo(0)
			.with(new JoinVertexWithVertexDegree<>())
				.setParallelism(parallelism)
				.name("Zero degree vertices");
	}

	return degree;
}
Example #18
Source File: ReduceITCase.java    From Flink-CEPplus with Apache License 2.0 5 votes vote down vote up 
@Test
public void testReduceOnTupleWithMultipleKeyExpressionsWithHashHint() throws Exception {
	/*
	 * Case 2 with String-based field expression
	 */

	final ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();

	DataSet<Tuple5<Integer, Long, Integer, String, Long>> ds = CollectionDataSets.get5TupleDataSet(env);
	DataSet<Tuple5<Integer, Long, Integer, String, Long>> reduceDs = ds
		.groupBy("f4", "f0").reduce(new Tuple5Reduce()).setCombineHint(CombineHint.HASH);

	List<Tuple5<Integer, Long, Integer, String, Long>> result = reduceDs
		.collect();

	String expected = "1,1,0,Hallo,1\n" +
		"2,3,2,Hallo Welt wie,1\n" +
		"2,2,1,Hallo Welt,2\n" +
		"3,9,0,P-),2\n" +
		"3,6,5,BCD,3\n" +
		"4,17,0,P-),1\n" +
		"4,17,0,P-),2\n" +
		"5,11,10,GHI,1\n" +
		"5,29,0,P-),2\n" +
		"5,25,0,P-),3\n";

	compareResultAsTuples(result, expected);
}
Example #19
Source File: EdgeMetrics.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public EdgeMetrics<K, VV, EV> run(Graph<K, VV, EV> input)
		throws Exception {
	super.run(input);

	// s, t, (d(s), d(t))
	DataSet<Edge<K, Tuple3<EV, LongValue, LongValue>>> edgeDegreePair = input
		.run(new EdgeDegreePair<K, VV, EV>()
			.setReduceOnTargetId(reduceOnTargetId)
			.setParallelism(parallelism));

	// s, d(s), count of (u, v) where deg(u) < deg(v) or (deg(u) == deg(v) and u < v)
	DataSet<Tuple3<K, LongValue, LongValue>> edgeStats = edgeDegreePair
		.map(new EdgeStats<>())
			.setParallelism(parallelism)
			.name("Edge stats")
		.groupBy(0)
		.reduce(new SumEdgeStats<>())
		.setCombineHint(CombineHint.HASH)
			.setParallelism(parallelism)
			.name("Sum edge stats");

	edgeMetricsHelper = new EdgeMetricsHelper<>();

	edgeStats
		.output(edgeMetricsHelper)
			.setParallelism(parallelism)
			.name("Edge metrics");

	return this;
}
Example #20
Source File: EdgeMetrics.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public EdgeMetrics<K, VV, EV> run(Graph<K, VV, EV> input)
		throws Exception {
	super.run(input);

	// s, t, (d(s), d(t))
	DataSet<Edge<K, Tuple3<EV, Degrees, Degrees>>> edgeDegreesPair = input
		.run(new EdgeDegreesPair<K, VV, EV>()
			.setParallelism(parallelism));

	// s, d(s), count of (u, v) where deg(u) < deg(v) or (deg(u) == deg(v) and u < v)
	DataSet<Tuple3<K, Degrees, LongValue>> edgeStats = edgeDegreesPair
		.flatMap(new EdgeStats<>())
			.setParallelism(parallelism)
			.name("Edge stats")
		.groupBy(0, 1)
		.reduceGroup(new ReduceEdgeStats<>())
			.setParallelism(parallelism)
			.name("Reduce edge stats")
		.groupBy(0)
		.reduce(new SumEdgeStats<>())
		.setCombineHint(CombineHint.HASH)
			.setParallelism(parallelism)
			.name("Sum edge stats");

	edgeMetricsHelper = new EdgeMetricsHelper<>();

	edgeStats
		.output(edgeMetricsHelper)
			.setParallelism(parallelism)
			.name("Edge metrics");

	return this;
}
Example #21
Source File: LocalClusteringCoefficient.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public DataSet<Result<K>> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// u, v, w, bitmask
	DataSet<TriangleListing.Result<K>> triangles = input
		.run(new TriangleListing<K, VV, EV>()
			.setParallelism(parallelism));

	// u, edge count
	DataSet<Tuple2<K, LongValue>> triangleVertices = triangles
		.flatMap(new SplitTriangles<>())
			.name("Split triangle vertices");

	// u, triangle count
	DataSet<Tuple2<K, LongValue>> vertexTriangleCount = triangleVertices
		.groupBy(0)
		.reduce(new CountTriangles<>())
		.setCombineHint(CombineHint.HASH)
			.name("Count triangles")
			.setParallelism(parallelism);

	// u, deg(u)
	DataSet<Vertex<K, Degrees>> vertexDegree = input
		.run(new VertexDegrees<K, VV, EV>()
			.setIncludeZeroDegreeVertices(includeZeroDegreeVertices.get())
			.setParallelism(parallelism));

	// u, deg(u), triangle count
	return vertexDegree
		.leftOuterJoin(vertexTriangleCount)
		.where(0)
		.equalTo(0)
		.with(new JoinVertexDegreeWithTriangleCount<>())
			.setParallelism(parallelism)
			.name("Clustering coefficient");
}
Example #22
Source File: LocalClusteringCoefficient.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public DataSet<Result<K>> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// u, v, w
	DataSet<TriangleListing.Result<K>> triangles = input
		.run(new TriangleListing<K, VV, EV>()
			.setParallelism(parallelism));

	// u, 1
	DataSet<Tuple2<K, LongValue>> triangleVertices = triangles
		.flatMap(new SplitTriangles<>())
			.name("Split triangle vertices");

	// u, triangle count
	DataSet<Tuple2<K, LongValue>> vertexTriangleCount = triangleVertices
		.groupBy(0)
		.reduce(new CountTriangles<>())
		.setCombineHint(CombineHint.HASH)
			.name("Count triangles")
			.setParallelism(parallelism);

	// u, deg(u)
	DataSet<Vertex<K, LongValue>> vertexDegree = input
		.run(new VertexDegree<K, VV, EV>()
			.setIncludeZeroDegreeVertices(includeZeroDegreeVertices.get())
			.setParallelism(parallelism));

	// u, deg(u), triangle count
	return vertexDegree
		.leftOuterJoin(vertexTriangleCount)
		.where(0)
		.equalTo(0)
		.with(new JoinVertexDegreeWithTriangleCount<>())
			.setParallelism(parallelism)
			.name("Clustering coefficient");
}
Example #23
Source File: EdgeMetrics.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public EdgeMetrics<K, VV, EV> run(Graph<K, VV, EV> input)
		throws Exception {
	super.run(input);

	// s, t, (d(s), d(t))
	DataSet<Edge<K, Tuple3<EV, Degrees, Degrees>>> edgeDegreesPair = input
		.run(new EdgeDegreesPair<K, VV, EV>()
			.setParallelism(parallelism));

	// s, d(s), count of (u, v) where deg(u) < deg(v) or (deg(u) == deg(v) and u < v)
	DataSet<Tuple3<K, Degrees, LongValue>> edgeStats = edgeDegreesPair
		.flatMap(new EdgeStats<>())
			.setParallelism(parallelism)
			.name("Edge stats")
		.groupBy(0, 1)
		.reduceGroup(new ReduceEdgeStats<>())
			.setParallelism(parallelism)
			.name("Reduce edge stats")
		.groupBy(0)
		.reduce(new SumEdgeStats<>())
		.setCombineHint(CombineHint.HASH)
			.setParallelism(parallelism)
			.name("Sum edge stats");

	edgeMetricsHelper = new EdgeMetricsHelper<>();

	edgeStats
		.output(edgeMetricsHelper)
			.setParallelism(parallelism)
			.name("Edge metrics");

	return this;
}
Example #24
Source File: EdgeMetrics.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public EdgeMetrics<K, VV, EV> run(Graph<K, VV, EV> input)
		throws Exception {
	super.run(input);

	// s, t, (d(s), d(t))
	DataSet<Edge<K, Tuple3<EV, LongValue, LongValue>>> edgeDegreePair = input
		.run(new EdgeDegreePair<K, VV, EV>()
			.setReduceOnTargetId(reduceOnTargetId)
			.setParallelism(parallelism));

	// s, d(s), count of (u, v) where deg(u) < deg(v) or (deg(u) == deg(v) and u < v)
	DataSet<Tuple3<K, LongValue, LongValue>> edgeStats = edgeDegreePair
		.map(new EdgeStats<>())
			.setParallelism(parallelism)
			.name("Edge stats")
		.groupBy(0)
		.reduce(new SumEdgeStats<>())
		.setCombineHint(CombineHint.HASH)
			.setParallelism(parallelism)
			.name("Sum edge stats");

	edgeMetricsHelper = new EdgeMetricsHelper<>();

	edgeStats
		.output(edgeMetricsHelper)
			.setParallelism(parallelism)
			.name("Edge metrics");

	return this;
}
Example #25
Source File: VertexInDegree.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public DataSet<Vertex<K, LongValue>> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// t
	DataSet<Vertex<K, LongValue>> targetIds = input
		.getEdges()
		.map(new MapEdgeToTargetId<>())
			.setParallelism(parallelism)
			.name("Edge to target ID");

	// t, d(t)
	DataSet<Vertex<K, LongValue>> targetDegree = targetIds
		.groupBy(0)
		.reduce(new DegreeCount<>())
		.setCombineHint(CombineHint.HASH)
			.setParallelism(parallelism)
			.name("Degree count");

	if (includeZeroDegreeVertices.get()) {
		targetDegree = input.getVertices()
			.leftOuterJoin(targetDegree)
			.where(0)
			.equalTo(0)
			.with(new JoinVertexWithVertexDegree<>())
				.setParallelism(parallelism)
				.name("Zero degree vertices");
	}

	return targetDegree;
}
Example #26
Source File: ReduceITCase.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Test
public void testReduceOnTupleWithMultipleKeyExpressionsWithHashHint() throws Exception {
	/*
	 * Case 2 with String-based field expression
	 */

	final ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();

	DataSet<Tuple5<Integer, Long, Integer, String, Long>> ds = CollectionDataSets.get5TupleDataSet(env);
	DataSet<Tuple5<Integer, Long, Integer, String, Long>> reduceDs = ds
		.groupBy("f4", "f0").reduce(new Tuple5Reduce()).setCombineHint(CombineHint.HASH);

	List<Tuple5<Integer, Long, Integer, String, Long>> result = reduceDs
		.collect();

	String expected = "1,1,0,Hallo,1\n" +
		"2,3,2,Hallo Welt wie,1\n" +
		"2,2,1,Hallo Welt,2\n" +
		"3,9,0,P-),2\n" +
		"3,6,5,BCD,3\n" +
		"4,17,0,P-),1\n" +
		"4,17,0,P-),2\n" +
		"5,11,10,GHI,1\n" +
		"5,29,0,P-),2\n" +
		"5,25,0,P-),3\n";

	compareResultAsTuples(result, expected);
}
Example #27
Source File: ReduceOperator.java    From flink with Apache License 2.0 5 votes vote down vote up 
public ReduceOperator(Grouping<IN> input, ReduceFunction<IN> function, String defaultName) {
	super(input.getInputDataSet(), input.getInputDataSet().getType());

	this.function = function;
	this.grouper = input;
	this.defaultName = defaultName;
	this.hint = CombineHint.OPTIMIZER_CHOOSES;
}
Example #28
Source File: LocalClusteringCoefficient.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public DataSet<Result<K>> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// u, v, w, bitmask
	DataSet<TriangleListing.Result<K>> triangles = input
		.run(new TriangleListing<K, VV, EV>()
			.setParallelism(parallelism));

	// u, edge count
	DataSet<Tuple2<K, LongValue>> triangleVertices = triangles
		.flatMap(new SplitTriangles<>())
			.name("Split triangle vertices");

	// u, triangle count
	DataSet<Tuple2<K, LongValue>> vertexTriangleCount = triangleVertices
		.groupBy(0)
		.reduce(new CountTriangles<>())
		.setCombineHint(CombineHint.HASH)
			.name("Count triangles")
			.setParallelism(parallelism);

	// u, deg(u)
	DataSet<Vertex<K, Degrees>> vertexDegree = input
		.run(new VertexDegrees<K, VV, EV>()
			.setIncludeZeroDegreeVertices(includeZeroDegreeVertices.get())
			.setParallelism(parallelism));

	// u, deg(u), triangle count
	return vertexDegree
		.leftOuterJoin(vertexTriangleCount)
		.where(0)
		.equalTo(0)
		.with(new JoinVertexDegreeWithTriangleCount<>())
			.setParallelism(parallelism)
			.name("Clustering coefficient");
}
Example #29
Source File: VertexDegree.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public DataSet<Vertex<K, LongValue>> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	MapFunction<Edge<K, EV>, Vertex<K, LongValue>> mapEdgeToId = reduceOnTargetId.get() ?
		new MapEdgeToTargetId<>() : new MapEdgeToSourceId<>();

	// v
	DataSet<Vertex<K, LongValue>> vertexIds = input
		.getEdges()
		.map(mapEdgeToId)
			.setParallelism(parallelism)
			.name("Edge to vertex ID");

	// v, deg(v)
	DataSet<Vertex<K, LongValue>> degree = vertexIds
		.groupBy(0)
		.reduce(new DegreeCount<>())
		.setCombineHint(CombineHint.HASH)
			.setParallelism(parallelism)
			.name("Degree count");

	if (includeZeroDegreeVertices.get()) {
		degree = input
			.getVertices()
			.leftOuterJoin(degree)
			.where(0)
			.equalTo(0)
			.with(new JoinVertexWithVertexDegree<>())
				.setParallelism(parallelism)
				.name("Zero degree vertices");
	}

	return degree;
}
Example #30
Source File: VertexOutDegree.java    From flink with Apache License 2.0 5 votes vote down vote up 
@Override
public DataSet<Vertex<K, LongValue>> runInternal(Graph<K, VV, EV> input)
		throws Exception {
	// s
	DataSet<Vertex<K, LongValue>> sourceIds = input
		.getEdges()
		.map(new MapEdgeToSourceId<>())
			.setParallelism(parallelism)
			.name("Edge to source ID");

	// s, d(s)
	DataSet<Vertex<K, LongValue>> sourceDegree = sourceIds
		.groupBy(0)
		.reduce(new DegreeCount<>())
		.setCombineHint(CombineHint.HASH)
			.setParallelism(parallelism)
			.name("Degree count");

	if (includeZeroDegreeVertices.get()) {
		sourceDegree = input.getVertices()
			.leftOuterJoin(sourceDegree)
			.where(0)
			.equalTo(0)
			.with(new JoinVertexWithVertexDegree<>())
				.setParallelism(parallelism)
				.name("Zero degree vertices");
	}

	return sourceDegree;
}