org.apache.flink.optimizer.plan.SinkPlanNode Java Examples

@Test
public void checkSinglePartitionedGroupedSource7() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	DataSource<Tuple3<Long, SomePojo, String>> data = env.fromCollection(tuple3PojoData, tuple3PojoType);

	data.getSplitDataProperties()
			.splitsPartitionedBy("f1.intField")
			.splitsGroupedBy("f1");

	data.output(new DiscardingOutputFormat<Tuple3<Long, SomePojo, String>>());

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	SourcePlanNode sourceNode = (SourcePlanNode) sinkNode.getPredecessor();

	GlobalProperties gprops = sourceNode.getGlobalProperties();
	LocalProperties lprops = sourceNode.getLocalProperties();

	Assert.assertTrue((new FieldSet(gprops.getPartitioningFields().toArray())).equals(new FieldSet(2)));
	Assert.assertTrue(gprops.getPartitioning() == PartitioningProperty.ANY_PARTITIONING);
	Assert.assertTrue(new FieldSet(lprops.getGroupedFields().toArray()).equals(new FieldSet(1,2,3)));
	Assert.assertTrue(lprops.getOrdering() == null);

}
 

public void dumpOptimizerPlanAsJSON(OptimizedPlan plan, PrintWriter writer) {
	Collection<SinkPlanNode> sinks = plan.getDataSinks();
	if (sinks instanceof List) {
		dumpOptimizerPlanAsJSON((List<SinkPlanNode>) sinks, writer);
	} else {
		List<SinkPlanNode> n = new ArrayList<SinkPlanNode>();
		n.addAll(sinks);
		dumpOptimizerPlanAsJSON(n, writer);
	}
}
 

@Test
public void testCoGroupWithTuples() {
	try {
		final Partitioner<Long> partitioner = new TestPartitionerLong();
		
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
		
		DataSet<Tuple2<Long, Long>> input1 = env.fromElements(new Tuple2<Long, Long>(0L, 0L));
		DataSet<Tuple3<Long, Long, Long>> input2 = env.fromElements(new Tuple3<Long, Long, Long>(0L, 0L, 0L));
		
		input1
			.coGroup(input2)
			.where(1).equalTo(0)
			.withPartitioner(partitioner)
			.with(new DummyCoGroupFunction<Tuple2<Long, Long>, Tuple3<Long, Long, Long>>())
			.output(new DiscardingOutputFormat<Tuple2<Tuple2<Long, Long>, Tuple3<Long, Long, Long>>>());
		
		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);
		
		SinkPlanNode sink = op.getDataSinks().iterator().next();
		DualInputPlanNode join = (DualInputPlanNode) sink.getInput().getSource();
		
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput1().getShipStrategy());
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput2().getShipStrategy());
		assertEquals(partitioner, join.getInput1().getPartitioner());
		assertEquals(partitioner, join.getInput2().getPartitioner());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

@Test
public void testCoGroupWithPojos() {
	try {
		final Partitioner<Integer> partitioner = new TestPartitionerInt();
		
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
		
		DataSet<Pojo2> input1 = env.fromElements(new Pojo2());
		DataSet<Pojo3> input2 = env.fromElements(new Pojo3());
		
		input1
			.coGroup(input2)
			.where("b").equalTo("a")
			.withPartitioner(partitioner)
			.with(new DummyCoGroupFunction<Pojo2, Pojo3>())
			.output(new DiscardingOutputFormat<Tuple2<Pojo2, Pojo3>>());
		
		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);
		
		SinkPlanNode sink = op.getDataSinks().iterator().next();
		DualInputPlanNode join = (DualInputPlanNode) sink.getInput().getSource();
		
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput1().getShipStrategy());
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput2().getShipStrategy());
		assertEquals(partitioner, join.getInput1().getPartitioner());
		assertEquals(partitioner, join.getInput2().getPartitioner());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

/**
 * Tests join program with replicated data source behind map partition.
 */
@Test
public void checkJoinWithReplicatedSourceInputBehindMapPartition() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	TupleTypeInfo<Tuple1<String>> typeInfo = TupleTypeInfo.getBasicTupleTypeInfo(String.class);
	ReplicatingInputFormat<Tuple1<String>, FileInputSplit> rif =
			new ReplicatingInputFormat<Tuple1<String>, FileInputSplit>(new TupleCsvInputFormat<Tuple1<String>>(new Path("/some/path"), typeInfo));

	DataSet<Tuple1<String>> source1 = env.createInput(rif, new TupleTypeInfo<Tuple1<String>>(BasicTypeInfo.STRING_TYPE_INFO));
	DataSet<Tuple1<String>> source2 = env.readCsvFile("/some/otherpath").types(String.class);

	DataSink<Tuple2<Tuple1<String>, Tuple1<String>>> out = source1
			.mapPartition(new IdPMap())
			.join(source2).where("*").equalTo("*")
			.writeAsText("/some/newpath");

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	// when join should have forward strategy on both sides
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	DualInputPlanNode joinNode = (DualInputPlanNode) sinkNode.getPredecessor();

	ShipStrategyType joinIn1 = joinNode.getInput1().getShipStrategy();
	ShipStrategyType joinIn2 = joinNode.getInput2().getShipStrategy();

	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, joinIn1);
	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, joinIn2);
}
 

/**
 * Creates a new plan finalizer.
 */
public PlanFinalizer() {
	this.allNodes = new HashSet<PlanNode>();
	this.sources = new ArrayList<SourcePlanNode>();
	this.sinks = new ArrayList<SinkPlanNode>();
	this.stackOfIterationNodes = new ArrayDeque<IterationPlanNode>();
}
 

/**
 * Simple Job: Map -> Reduce -> Map -> Reduce. All functions preserve all fields (hence all properties).
 * 
 * Increases parallelism between 2nd map and 2nd reduce, so the hash partitioning from 1st reduce is not reusable.
 * Expected to re-establish partitioning between map and reduce (hash).
 */
@Test
public void checkPropertyHandlingWithIncreasingGlobalParallelism2() {
	final int p = DEFAULT_PARALLELISM;

	// construct the plan
	ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
	env.setParallelism(p);
	DataSet<Long> set1 = env.generateSequence(0,1).setParallelism(p);

	set1.map(new IdentityMapper<Long>())
			.withForwardedFields("*").setParallelism(p).name("Map1")
			.groupBy("*").reduceGroup(new IdentityGroupReducer<Long>())
			.withForwardedFields("*").setParallelism(p).name("Reduce1")
			.map(new IdentityMapper<Long>())
			.withForwardedFields("*").setParallelism(p).name("Map2")
			.groupBy("*").reduceGroup(new IdentityGroupReducer<Long>())
			.withForwardedFields("*").setParallelism(p * 2).name("Reduce2")
			.output(new DiscardingOutputFormat<Long>()).setParallelism(p * 2).name("Sink");

	Plan plan = env.createProgramPlan();
	
	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);
	
	// check the optimized Plan
	// when reducer 1 distributes its data across the instances of map2, it needs to employ a local hash method,
	// because map2 has twice as many instances and key/value pairs with the same key need to be processed by the same
	// mapper respectively reducer
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	SingleInputPlanNode red2Node = (SingleInputPlanNode) sinkNode.getPredecessor();
	SingleInputPlanNode map2Node = (SingleInputPlanNode) red2Node.getPredecessor();
	
	ShipStrategyType mapIn = map2Node.getInput().getShipStrategy();
	ShipStrategyType reduceIn = red2Node.getInput().getShipStrategy();
	
	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, mapIn);
	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.PARTITION_HASH, reduceIn);
}
 

/**
 * Simple Job: Map -> Reduce -> Map -> Reduce. All functions preserve all fields (hence all properties).
 * 
 * Increases parallelism between 1st reduce and 2nd map, so the hash partitioning from 1st reduce is not reusable.
 * Expected to re-establish partitioning between reduce and map, via hash, because random is a full network
 * transit as well.
 */
@Test
public void checkPropertyHandlingWithIncreasingGlobalParallelism1() {
	final int p = DEFAULT_PARALLELISM;

	// construct the plan
	ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
	env.setParallelism(p);
	DataSet<Long> set1 = env.generateSequence(0,1).setParallelism(p);

	set1.map(new IdentityMapper<Long>())
				.withForwardedFields("*").setParallelism(p).name("Map1")
			.groupBy("*").reduceGroup(new IdentityGroupReducer<Long>())
				.withForwardedFields("*").setParallelism(p).name("Reduce1")
			.map(new IdentityMapper<Long>())
				.withForwardedFields("*").setParallelism(p * 2).name("Map2")
			.groupBy("*").reduceGroup(new IdentityGroupReducer<Long>())
				.withForwardedFields("*").setParallelism(p * 2).name("Reduce2")
			.output(new DiscardingOutputFormat<Long>()).setParallelism(p * 2).name("Sink");

	Plan plan = env.createProgramPlan();
	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);
	
	// check the optimized Plan
	// when reducer 1 distributes its data across the instances of map2, it needs to employ a local hash method,
	// because map2 has twice as many instances and key/value pairs with the same key need to be processed by the same
	// mapper respectively reducer
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	SingleInputPlanNode red2Node = (SingleInputPlanNode) sinkNode.getPredecessor();
	SingleInputPlanNode map2Node = (SingleInputPlanNode) red2Node.getPredecessor();
	
	ShipStrategyType mapIn = map2Node.getInput().getShipStrategy();
	ShipStrategyType redIn = red2Node.getInput().getShipStrategy();
	
	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.PARTITION_HASH, mapIn);
	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, redIn);
}
 

@Test
public void testDistinctPreservesPartitioningOfDistinctFields() {
	try {
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
		env.setParallelism(4);
		
		@SuppressWarnings("unchecked")
		DataSet<Tuple2<Long, Long>> data = env.fromElements(new Tuple2<Long, Long>(0L, 0L), new Tuple2<Long, Long>(1L, 1L))
				.map(new IdentityMapper<Tuple2<Long,Long>>()).setParallelism(4);
		
		data.distinct(0)
			.groupBy(0)
			.sum(1)
			.output(new DiscardingOutputFormat<Tuple2<Long, Long>>());
		
		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);
		
		SinkPlanNode sink = op.getDataSinks().iterator().next();
		SingleInputPlanNode reducer = (SingleInputPlanNode) sink.getInput().getSource();
		SingleInputPlanNode distinctReducer = (SingleInputPlanNode) reducer.getInput().getSource();
		
		assertEquals(ShipStrategyType.FORWARD, sink.getInput().getShipStrategy());
		
		// reducer can be forward, reuses partitioning from distinct
		assertEquals(ShipStrategyType.FORWARD, reducer.getInput().getShipStrategy());
		
		// distinct reducer is partitioned
		assertEquals(ShipStrategyType.PARTITION_HASH, distinctReducer.getInput().getShipStrategy());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

@Test
public void checkSinglePartitionedOrderedSource6() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	DataSource<Tuple3<Long, SomePojo, String>> data = env.fromCollection(tuple3PojoData, tuple3PojoType);

	data.getSplitDataProperties()
			.splitsPartitionedBy("f1.intField")
			.splitsOrderedBy("f1", new Order[]{Order.DESCENDING});

	data.output(new DiscardingOutputFormat<Tuple3<Long, SomePojo, String>>());

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	SourcePlanNode sourceNode = (SourcePlanNode) sinkNode.getPredecessor();

	GlobalProperties gprops = sourceNode.getGlobalProperties();
	LocalProperties lprops = sourceNode.getLocalProperties();

	Assert.assertTrue((new FieldSet(gprops.getPartitioningFields().toArray())).equals(new FieldSet(2)));
	Assert.assertTrue(gprops.getPartitioning() == PartitioningProperty.ANY_PARTITIONING);
	Assert.assertTrue(new FieldSet(lprops.getGroupedFields().toArray()).equals(new FieldSet(1,2,3)));
	Assert.assertTrue(lprops.getOrdering() == null);

}
 

@Test
public void testCoGroupWithKeySelectors() {
	try {
		final Partitioner<Integer> partitioner = new TestPartitionerInt();
		
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
		
		DataSet<Pojo2> input1 = env.fromElements(new Pojo2());
		DataSet<Pojo3> input2 = env.fromElements(new Pojo3());
		
		input1
			.coGroup(input2)
			.where(new Pojo2KeySelector()).equalTo(new Pojo3KeySelector())
			.withPartitioner(partitioner)
			.with(new DummyCoGroupFunction<Pojo2, Pojo3>())
			.output(new DiscardingOutputFormat<Tuple2<Pojo2, Pojo3>>());
		
		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);
		
		SinkPlanNode sink = op.getDataSinks().iterator().next();
		DualInputPlanNode join = (DualInputPlanNode) sink.getInput().getSource();
		
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput1().getShipStrategy());
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput2().getShipStrategy());
		assertEquals(partitioner, join.getInput1().getPartitioner());
		assertEquals(partitioner, join.getInput2().getPartitioner());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

private void checkStandardStrategies(SingleInputPlanNode map, DualInputPlanNode join, SingleInputPlanNode combiner,
		SingleInputPlanNode reducer, SinkPlanNode sink) {
	// check ship strategies that are always fix
	Assert.assertEquals(ShipStrategyType.FORWARD, map.getInput().getShipStrategy());
	Assert.assertEquals(ShipStrategyType.FORWARD, sink.getInput().getShipStrategy());

	// check the driver strategies that are always fix
	Assert.assertEquals(DriverStrategy.FLAT_MAP, map.getDriverStrategy());
	Assert.assertEquals(DriverStrategy.SORTED_GROUP_REDUCE, reducer.getDriverStrategy());
	Assert.assertEquals(DriverStrategy.NONE, sink.getDriverStrategy());
	if (combiner != null) {
		Assert.assertEquals(DriverStrategy.SORTED_GROUP_COMBINE, combiner.getDriverStrategy());
		Assert.assertEquals(LocalStrategy.NONE, combiner.getInput().getLocalStrategy());
	}
}
 

@Test
public void testCoGroupWithPojos() {
	try {
		final Partitioner<Integer> partitioner = new TestPartitionerInt();
		
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
		
		DataSet<Pojo2> input1 = env.fromElements(new Pojo2());
		DataSet<Pojo3> input2 = env.fromElements(new Pojo3());
		
		input1
			.coGroup(input2)
			.where("b").equalTo("a")
			.withPartitioner(partitioner)
			.with(new DummyCoGroupFunction<Pojo2, Pojo3>())
			.output(new DiscardingOutputFormat<Tuple2<Pojo2, Pojo3>>());
		
		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);
		
		SinkPlanNode sink = op.getDataSinks().iterator().next();
		DualInputPlanNode join = (DualInputPlanNode) sink.getInput().getSource();
		
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput1().getShipStrategy());
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput2().getShipStrategy());
		assertEquals(partitioner, join.getInput1().getPartitioner());
		assertEquals(partitioner, join.getInput2().getPartitioner());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

@Test
public void testCustomPartitioningTupleGroupReduceSorted2() {
	try {
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();

		DataSet<Pojo4> data = env.fromElements(new Pojo4())
				.rebalance().setParallelism(4);

		data.groupBy("a").withPartitioner(new TestPartitionerInt())
				.sortGroup("b", Order.ASCENDING)
				.sortGroup("c", Order.DESCENDING)
				.reduceGroup(new IdentityGroupReducerCombinable<Pojo4>())
				.output(new DiscardingOutputFormat<Pojo4>());

		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);

		SinkPlanNode sink = op.getDataSinks().iterator().next();
		SingleInputPlanNode reducer = (SingleInputPlanNode) sink.getInput().getSource();
		SingleInputPlanNode combiner = (SingleInputPlanNode) reducer.getInput().getSource();

		assertEquals(ShipStrategyType.FORWARD, sink.getInput().getShipStrategy());
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, reducer.getInput().getShipStrategy());
		assertEquals(ShipStrategyType.FORWARD, combiner.getInput().getShipStrategy());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

private SinkPlanNode findSink(Collection<SinkPlanNode> collection, String name) {
	for (SinkPlanNode node : collection) {
		String nodeName = node.getOptimizerNode().getOperator().getName();
		if (nodeName != null && nodeName.equals(name)) {
			return node;
		}
	}

	throw new IllegalArgumentException("No node with that name was found.");
}
 

@Test
public void checkSinglePartitionedGroupedSource3() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	DataSource<Tuple2<Long, String>> data =
			env.readCsvFile("/some/path").types(Long.class, String.class);

	data.getSplitDataProperties()
			.splitsPartitionedBy(1)
			.splitsGroupedBy(0);

	data.output(new DiscardingOutputFormat<Tuple2<Long, String>>());

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	SourcePlanNode sourceNode = (SourcePlanNode) sinkNode.getPredecessor();

	GlobalProperties gprops = sourceNode.getGlobalProperties();
	LocalProperties lprops = sourceNode.getLocalProperties();

	Assert.assertTrue((new FieldSet(gprops.getPartitioningFields().toArray())).equals(new FieldSet(1)));
	Assert.assertTrue(gprops.getPartitioning() == PartitioningProperty.ANY_PARTITIONING);
	Assert.assertTrue(lprops.getGroupedFields() == null);
	Assert.assertTrue(lprops.getOrdering() == null);

}
 

@Test
public void testCustomPartitioningTupleGroupReduce() {
	try {
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();

		DataSet<Pojo2> data = env.fromElements(new Pojo2())
				.rebalance().setParallelism(4);

		data.groupBy("a").withPartitioner(new TestPartitionerInt())
				.reduceGroup(new IdentityGroupReducerCombinable<Pojo2>())
				.output(new DiscardingOutputFormat<Pojo2>());

		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);

		SinkPlanNode sink = op.getDataSinks().iterator().next();
		SingleInputPlanNode reducer = (SingleInputPlanNode) sink.getInput().getSource();
		SingleInputPlanNode combiner = (SingleInputPlanNode) reducer.getInput().getSource();

		assertEquals(ShipStrategyType.FORWARD, sink.getInput().getShipStrategy());
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, reducer.getInput().getShipStrategy());
		assertEquals(ShipStrategyType.FORWARD, combiner.getInput().getShipStrategy());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

private SinkPlanNode findSink(Collection<SinkPlanNode> collection, String name) {
	for (SinkPlanNode node : collection) {
		String nodeName = node.getOptimizerNode().getOperator().getName();
		if (nodeName != null && nodeName.equals(name)) {
			return node;
		}
	}

	throw new IllegalArgumentException("No node with that name was found.");
}
 

public void dumpOptimizerPlanAsJSON(OptimizedPlan plan, PrintWriter writer) {
	Collection<SinkPlanNode> sinks = plan.getDataSinks();
	if (sinks instanceof List) {
		dumpOptimizerPlanAsJSON((List<SinkPlanNode>) sinks, writer);
	} else {
		List<SinkPlanNode> n = new ArrayList<SinkPlanNode>();
		n.addAll(sinks);
		dumpOptimizerPlanAsJSON(n, writer);
	}
}
 

@Test
public void testPipelineBreakerWithBroadcastVariable() {
	try {
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
		env.getConfig().setExecutionMode(ExecutionMode.PIPELINED);
		env.setParallelism(64);
		
		DataSet<Long> source = env.generateSequence(1, 10).map(new IdentityMapper<Long>());
		
		DataSet<Long> result = source.map(new IdentityMapper<Long>())
									.map(new IdentityMapper<Long>())
										.withBroadcastSet(source, "bc");
		
		result.output(new DiscardingOutputFormat<Long>());
		
		
		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);
		
		SinkPlanNode sink = op.getDataSinks().iterator().next();
		SingleInputPlanNode mapper = (SingleInputPlanNode) sink.getInput().getSource();
		SingleInputPlanNode mapperInput = (SingleInputPlanNode) mapper.getInput().getSource();
		
		assertEquals(TempMode.NONE, mapper.getInput().getTempMode());
		assertEquals(TempMode.NONE, mapper.getBroadcastInputs().get(0).getTempMode());
		
		assertEquals(DataExchangeMode.BATCH, mapperInput.getInput().getDataExchangeMode());
		assertEquals(DataExchangeMode.BATCH, mapper.getBroadcastInputs().get(0).getDataExchangeMode());
		
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

@Test
public void testCustomPartitioningTupleGroupReduce() {
	try {
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();

		DataSet<Pojo2> data = env.fromElements(new Pojo2())
				.rebalance().setParallelism(4);

		data.groupBy("a").withPartitioner(new TestPartitionerInt())
				.reduceGroup(new IdentityGroupReducerCombinable<Pojo2>())
				.output(new DiscardingOutputFormat<Pojo2>());

		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);

		SinkPlanNode sink = op.getDataSinks().iterator().next();
		SingleInputPlanNode reducer = (SingleInputPlanNode) sink.getInput().getSource();
		SingleInputPlanNode combiner = (SingleInputPlanNode) reducer.getInput().getSource();

		assertEquals(ShipStrategyType.FORWARD, sink.getInput().getShipStrategy());
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, reducer.getInput().getShipStrategy());
		assertEquals(ShipStrategyType.FORWARD, combiner.getInput().getShipStrategy());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

@Test
public void testPartitioningReused() {
	try {
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
		
		@SuppressWarnings("unchecked")
		DataSet<Tuple3<Long, Long, Long>> input = env.fromElements(new Tuple3<Long, Long, Long>(0L, 0L, 0L));
		
		input
			.groupBy(0).sum(1)
			.groupBy(0, 1).sum(2)
			.output(new DiscardingOutputFormat<Tuple3<Long, Long, Long>>());
		
		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);
		
		SinkPlanNode sink = op.getDataSinks().iterator().next();
		
		SingleInputPlanNode agg2Reducer = (SingleInputPlanNode) sink.getInput().getSource();
		SingleInputPlanNode agg1Reducer = (SingleInputPlanNode) agg2Reducer.getInput().getSource();
		
		assertEquals(ShipStrategyType.FORWARD, agg2Reducer.getInput().getShipStrategy());
		
		assertEquals(ShipStrategyType.PARTITION_HASH, agg1Reducer.getInput().getShipStrategy());
		assertEquals(new FieldList(0), agg1Reducer.getInput().getShipStrategyKeys());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

@Override
public void postPass(OptimizedPlan plan) {

	executionConfig = plan.getOriginalPlan().getExecutionConfig();

	for (SinkPlanNode sink : plan.getDataSinks()) {
		traverse(sink);
	}
}
 

@Test
public void checkSinglePartitionedOrderedSource2() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	DataSource<Tuple2<Long, String>> data =
			env.readCsvFile("/some/path").types(Long.class, String.class);

	data.getSplitDataProperties()
			.splitsPartitionedBy(1)
			.splitsOrderedBy(new int[]{1, 0}, new Order[]{Order.ASCENDING, Order.DESCENDING});

	data.output(new DiscardingOutputFormat<Tuple2<Long, String>>());

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	SourcePlanNode sourceNode = (SourcePlanNode) sinkNode.getPredecessor();

	GlobalProperties gprops = sourceNode.getGlobalProperties();
	LocalProperties lprops = sourceNode.getLocalProperties();

	Assert.assertTrue((new FieldSet(gprops.getPartitioningFields().toArray())).equals(new FieldSet(1)));
	Assert.assertTrue(gprops.getPartitioning() == PartitioningProperty.ANY_PARTITIONING);
	Assert.assertTrue((new FieldSet(lprops.getGroupedFields().toArray())).equals(new FieldSet(1, 0)));
	Assert.assertTrue(lprops.getOrdering() == null);

}
 

@Test
public void testCoGroupWithTuples() {
	try {
		final Partitioner<Long> partitioner = new TestPartitionerLong();
		
		ExecutionEnvironment env = ExecutionEnvironment.getExecutionEnvironment();
		
		DataSet<Tuple2<Long, Long>> input1 = env.fromElements(new Tuple2<Long, Long>(0L, 0L));
		DataSet<Tuple3<Long, Long, Long>> input2 = env.fromElements(new Tuple3<Long, Long, Long>(0L, 0L, 0L));
		
		input1
			.coGroup(input2)
			.where(1).equalTo(0)
			.withPartitioner(partitioner)
			.with(new DummyCoGroupFunction<Tuple2<Long, Long>, Tuple3<Long, Long, Long>>())
			.output(new DiscardingOutputFormat<Tuple2<Tuple2<Long, Long>, Tuple3<Long, Long, Long>>>());
		
		Plan p = env.createProgramPlan();
		OptimizedPlan op = compileNoStats(p);
		
		SinkPlanNode sink = op.getDataSinks().iterator().next();
		DualInputPlanNode join = (DualInputPlanNode) sink.getInput().getSource();
		
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput1().getShipStrategy());
		assertEquals(ShipStrategyType.PARTITION_CUSTOM, join.getInput2().getShipStrategy());
		assertEquals(partitioner, join.getInput1().getPartitioner());
		assertEquals(partitioner, join.getInput2().getPartitioner());
	}
	catch (Exception e) {
		e.printStackTrace();
		fail(e.getMessage());
	}
}
 

@Test
public void checkSinglePartitionedGroupedSource7() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	DataSource<Tuple3<Long, SomePojo, String>> data = env.fromCollection(tuple3PojoData, tuple3PojoType);

	data.getSplitDataProperties()
			.splitsPartitionedBy("f1.intField")
			.splitsGroupedBy("f1");

	data.output(new DiscardingOutputFormat<Tuple3<Long, SomePojo, String>>());

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	SourcePlanNode sourceNode = (SourcePlanNode) sinkNode.getPredecessor();

	GlobalProperties gprops = sourceNode.getGlobalProperties();
	LocalProperties lprops = sourceNode.getLocalProperties();

	Assert.assertTrue((new FieldSet(gprops.getPartitioningFields().toArray())).equals(new FieldSet(2)));
	Assert.assertTrue(gprops.getPartitioning() == PartitioningProperty.ANY_PARTITIONING);
	Assert.assertTrue(new FieldSet(lprops.getGroupedFields().toArray()).equals(new FieldSet(1,2,3)));
	Assert.assertTrue(lprops.getOrdering() == null);

}
 

/**
 * Tests join program with replicated data source behind filter.
 */
@Test
public void checkJoinWithReplicatedSourceInputBehindFilter() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	TupleTypeInfo<Tuple1<String>> typeInfo = TupleTypeInfo.getBasicTupleTypeInfo(String.class);
	ReplicatingInputFormat<Tuple1<String>, FileInputSplit> rif =
			new ReplicatingInputFormat<Tuple1<String>, FileInputSplit>(new TupleCsvInputFormat<Tuple1<String>>(new Path("/some/path"), typeInfo));

	DataSet<Tuple1<String>> source1 = env.createInput(rif, new TupleTypeInfo<Tuple1<String>>(BasicTypeInfo.STRING_TYPE_INFO));
	DataSet<Tuple1<String>> source2 = env.readCsvFile("/some/otherpath").types(String.class);

	DataSink<Tuple2<Tuple1<String>, Tuple1<String>>> out = source1
			.filter(new NoFilter())
			.join(source2).where("*").equalTo("*")
			.writeAsText("/some/newpath");

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	// when join should have forward strategy on both sides
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	DualInputPlanNode joinNode = (DualInputPlanNode) sinkNode.getPredecessor();

	ShipStrategyType joinIn1 = joinNode.getInput1().getShipStrategy();
	ShipStrategyType joinIn2 = joinNode.getInput2().getShipStrategy();

	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, joinIn1);
	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, joinIn2);
}
 

/**
 * Tests join program with replicated data source behind flatMap.
 */
@Test
public void checkJoinWithReplicatedSourceInputBehindFlatMap() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	TupleTypeInfo<Tuple1<String>> typeInfo = TupleTypeInfo.getBasicTupleTypeInfo(String.class);
	ReplicatingInputFormat<Tuple1<String>, FileInputSplit> rif =
			new ReplicatingInputFormat<Tuple1<String>, FileInputSplit>(new TupleCsvInputFormat<Tuple1<String>>(new Path("/some/path"), typeInfo));

	DataSet<Tuple1<String>> source1 = env.createInput(rif, new TupleTypeInfo<Tuple1<String>>(BasicTypeInfo.STRING_TYPE_INFO));
	DataSet<Tuple1<String>> source2 = env.readCsvFile("/some/otherpath").types(String.class);

	DataSink<Tuple2<Tuple1<String>, Tuple1<String>>> out = source1
			.flatMap(new IdFlatMap())
			.join(source2).where("*").equalTo("*")
			.writeAsText("/some/newpath");

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	// when join should have forward strategy on both sides
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	DualInputPlanNode joinNode = (DualInputPlanNode) sinkNode.getPredecessor();

	ShipStrategyType joinIn1 = joinNode.getInput1().getShipStrategy();
	ShipStrategyType joinIn2 = joinNode.getInput2().getShipStrategy();

	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, joinIn1);
	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, joinIn2);
}
 

/**
 * Tests join program with replicated data source behind map partition.
 */
@Test
public void checkJoinWithReplicatedSourceInputBehindMapPartition() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	TupleTypeInfo<Tuple1<String>> typeInfo = TupleTypeInfo.getBasicTupleTypeInfo(String.class);
	ReplicatingInputFormat<Tuple1<String>, FileInputSplit> rif =
			new ReplicatingInputFormat<Tuple1<String>, FileInputSplit>(new TupleCsvInputFormat<Tuple1<String>>(new Path("/some/path"), typeInfo));

	DataSet<Tuple1<String>> source1 = env.createInput(rif, new TupleTypeInfo<Tuple1<String>>(BasicTypeInfo.STRING_TYPE_INFO));
	DataSet<Tuple1<String>> source2 = env.readCsvFile("/some/otherpath").types(String.class);

	DataSink<Tuple2<Tuple1<String>, Tuple1<String>>> out = source1
			.mapPartition(new IdPMap())
			.join(source2).where("*").equalTo("*")
			.writeAsText("/some/newpath");

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	// when join should have forward strategy on both sides
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	DualInputPlanNode joinNode = (DualInputPlanNode) sinkNode.getPredecessor();

	ShipStrategyType joinIn1 = joinNode.getInput1().getShipStrategy();
	ShipStrategyType joinIn2 = joinNode.getInput2().getShipStrategy();

	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, joinIn1);
	Assert.assertEquals("Invalid ship strategy for an operator.", ShipStrategyType.FORWARD, joinIn2);
}
 

@Test
public void checkSinglePartitionedGroupedSource8() {

	ExecutionEnvironment env = ExecutionEnvironment.createLocalEnvironment();
	env.setParallelism(DEFAULT_PARALLELISM);

	DataSource<Tuple3<Long, SomePojo, String>> data = env.fromCollection(tuple3PojoData, tuple3PojoType);

	data.getSplitDataProperties()
			.splitsPartitionedBy("f1")
			.splitsGroupedBy("f1.stringField");

	data.output(new DiscardingOutputFormat<Tuple3<Long, SomePojo, String>>());

	Plan plan = env.createProgramPlan();

	// submit the plan to the compiler
	OptimizedPlan oPlan = compileNoStats(plan);

	// check the optimized Plan
	SinkPlanNode sinkNode = oPlan.getDataSinks().iterator().next();
	SourcePlanNode sourceNode = (SourcePlanNode) sinkNode.getPredecessor();

	GlobalProperties gprops = sourceNode.getGlobalProperties();
	LocalProperties lprops = sourceNode.getLocalProperties();

	Assert.assertTrue((new FieldSet(gprops.getPartitioningFields().toArray())).equals(new FieldSet(1,2,3)));
	Assert.assertTrue(gprops.getPartitioning() == PartitioningProperty.ANY_PARTITIONING);
	Assert.assertTrue(lprops.getGroupedFields() == null);
	Assert.assertTrue(lprops.getOrdering() == null);

}