Examples with MergeIterator org.apache.flink.runtime.operators.sort.MergeIterator used on opensource projects

Example 1 with MergeIterator

use of org.apache.flink.runtime.operators.sort.MergeIterator in project flink by apache.

the class RandomSortMergeInnerJoinTest method testMergeWithHighNumberOfCommonKeys.

@Test
public void testMergeWithHighNumberOfCommonKeys() {
    // the size of the left and right inputs
    final int input1Size = 200;
    final int input2Size = 100;
    final int input1Duplicates = 10;
    final int input2Duplicates = 4000;
    final int duplicateKey = 13;
    try {
        final TupleGenerator generator1 = new TupleGenerator(SEED1, 500, 4096, KeyMode.SORTED, ValueMode.RANDOM_LENGTH);
        final TupleGenerator generator2 = new TupleGenerator(SEED2, 500, 2048, KeyMode.SORTED, ValueMode.RANDOM_LENGTH);
        final TestData.TupleGeneratorIterator gen1Iter = new TestData.TupleGeneratorIterator(generator1, input1Size);
        final TestData.TupleGeneratorIterator gen2Iter = new TestData.TupleGeneratorIterator(generator2, input2Size);
        final TestData.TupleConstantValueIterator const1Iter = new TestData.TupleConstantValueIterator(duplicateKey, "LEFT String for Duplicate Keys", input1Duplicates);
        final TestData.TupleConstantValueIterator const2Iter = new TestData.TupleConstantValueIterator(duplicateKey, "RIGHT String for Duplicate Keys", input2Duplicates);
        final List<MutableObjectIterator<Tuple2<Integer, String>>> inList1 = new ArrayList<>();
        inList1.add(gen1Iter);
        inList1.add(const1Iter);
        final List<MutableObjectIterator<Tuple2<Integer, String>>> inList2 = new ArrayList<>();
        inList2.add(gen2Iter);
        inList2.add(const2Iter);
        MutableObjectIterator<Tuple2<Integer, String>> input1 = new MergeIterator<>(inList1, comparator1.duplicate());
        MutableObjectIterator<Tuple2<Integer, String>> input2 = new MergeIterator<>(inList2, comparator2.duplicate());
        // collect expected data
        final Map<Integer, Collection<Match>> expectedMatchesMap = matchValues(collectData(input1), collectData(input2));
        // re-create the whole thing for actual processing
        // reset the generators and iterators
        generator1.reset();
        generator2.reset();
        const1Iter.reset();
        const2Iter.reset();
        gen1Iter.reset();
        gen2Iter.reset();
        inList1.clear();
        inList1.add(gen1Iter);
        inList1.add(const1Iter);
        inList2.clear();
        inList2.add(gen2Iter);
        inList2.add(const2Iter);
        input1 = new MergeIterator<>(inList1, comparator1.duplicate());
        input2 = new MergeIterator<>(inList2, comparator2.duplicate());
        StreamOperator operator = getOperator();
        match(expectedMatchesMap, transformToBinary(join(operator, input1, input2)));
        // assert that each expected match was seen
        for (Map.Entry<Integer, Collection<Match>> entry : expectedMatchesMap.entrySet()) {
            if (!entry.getValue().isEmpty()) {
                Assert.fail("Collection for key " + entry.getKey() + " is not empty");
            }
        }
    } catch (Exception e) {
        e.printStackTrace();
        Assert.fail("An exception occurred during the test: " + e.getMessage());
    }
}

Example 2 with MergeIterator

use of org.apache.flink.runtime.operators.sort.MergeIterator in project flink by apache.

the class RandomSortMergeOuterJoinTest method testOuterJoinWithHighNumberOfCommonKeys.

@SuppressWarnings("unchecked, rawtypes")
protected void testOuterJoinWithHighNumberOfCommonKeys(FlinkJoinType outerJoinType, int input1Size, int input1Duplicates, int input1ValueLength, float input1KeyDensity, int input2Size, int input2Duplicates, int input2ValueLength, float input2KeyDensity) {
    TypeComparator<Tuple2<Integer, String>> comparator1 = new TupleComparator<>(new int[] { 0 }, new TypeComparator<?>[] { new IntComparator(true) }, new TypeSerializer<?>[] { IntSerializer.INSTANCE });
    TypeComparator<Tuple2<Integer, String>> comparator2 = new TupleComparator<>(new int[] { 0 }, new TypeComparator<?>[] { new IntComparator(true) }, new TypeSerializer<?>[] { IntSerializer.INSTANCE });
    final int duplicateKey = 13;
    try {
        final TupleGenerator generator1 = new TupleGenerator(SEED1, 500, input1KeyDensity, input1ValueLength, KeyMode.SORTED_SPARSE, ValueMode.RANDOM_LENGTH, null);
        final TupleGenerator generator2 = new TupleGenerator(SEED2, 500, input2KeyDensity, input2ValueLength, KeyMode.SORTED_SPARSE, ValueMode.RANDOM_LENGTH, null);
        final TupleGeneratorIterator gen1Iter = new TupleGeneratorIterator(generator1, input1Size);
        final TupleGeneratorIterator gen2Iter = new TupleGeneratorIterator(generator2, input2Size);
        final TupleConstantValueIterator const1Iter = new TupleConstantValueIterator(duplicateKey, "LEFT String for Duplicate Keys", input1Duplicates);
        final TupleConstantValueIterator const2Iter = new TupleConstantValueIterator(duplicateKey, "RIGHT String for Duplicate Keys", input2Duplicates);
        final List<MutableObjectIterator<Tuple2<Integer, String>>> inList1 = new ArrayList<>();
        inList1.add(gen1Iter);
        inList1.add(const1Iter);
        final List<MutableObjectIterator<Tuple2<Integer, String>>> inList2 = new ArrayList<>();
        inList2.add(gen2Iter);
        inList2.add(const2Iter);
        MutableObjectIterator<Tuple2<Integer, String>> input1 = new MergeIterator<>(inList1, comparator1.duplicate());
        MutableObjectIterator<Tuple2<Integer, String>> input2 = new MergeIterator<>(inList2, comparator2.duplicate());
        // collect expected data
        final Map<Integer, Collection<Match>> expectedMatchesMap = joinValues(RandomSortMergeInnerJoinTest.collectData(input1), RandomSortMergeInnerJoinTest.collectData(input2), outerJoinType);
        // re-create the whole thing for actual processing
        // reset the generators and iterators
        generator1.reset();
        generator2.reset();
        const1Iter.reset();
        const2Iter.reset();
        gen1Iter.reset();
        gen2Iter.reset();
        inList1.clear();
        inList1.add(gen1Iter);
        inList1.add(const1Iter);
        inList2.clear();
        inList2.add(gen2Iter);
        inList2.add(const2Iter);
        input1 = new MergeIterator<>(inList1, comparator1.duplicate());
        input2 = new MergeIterator<>(inList2, comparator2.duplicate());
        StreamOperator operator = getOperator(outerJoinType);
        RandomSortMergeInnerJoinTest.match(expectedMatchesMap, RandomSortMergeInnerJoinTest.transformToBinary(myJoin(operator, input1, input2)));
        // assert that each expected match was seen
        for (Entry<Integer, Collection<Match>> entry : expectedMatchesMap.entrySet()) {
            if (!entry.getValue().isEmpty()) {
                Assert.fail("Collection for key " + entry.getKey() + " is not empty");
            }
        }
    } catch (Exception e) {
        e.printStackTrace();
        Assert.fail("An exception occurred during the test: " + e.getMessage());
    }
}