如何使用流从两个列表或数组乘法中查找元素对

Question

I have two lists of numbers and I'd like to find all possible pairs of numbers. For example, given the lists [1, 2, 3] and [3, 4] the result should be:

[(1, 3), (1, 4), (2, 3), (2, 4), (3, 3), (3, 4)]

I know I can do this using a for loop but is there any more concise way to do it using Java 8 streams ?

I tried the following, but I'm missing something as I'm getting List<Stream<int[]>> instead of List<int[]> .

public static void main(String[] args) {
    List<Integer> list1 = Arrays.asList(1, 2, 3);
    List<Integer> list2 = Arrays.asList(3, 4);
    List<int[]> pairs = list1.stream()
                             .map(i -> list2.stream()
                                            .map(j -> new int[]{i, j}))
                             .collect(Collectors.toList());
    pairs.forEach(i -> {
        System.out.println("{" + i[0] + "," + i[1] + "}");
    });
}

Answer 1

Use flatMap() method instead of map() , it will combine the streams into one. Refer : Difference Between map() and flatMap() and flatMap() example

Answer 2

你只需要用flatMap()替换你的第一个map() flatMap() 。

Answer 3

Here is a solution using IntStream with two int arrays as the source instead of List<Integer> . I wanted to see if it was possible to solve this problem without boxing every int as an Integer .

int[] one = new int[]{1, 2, 3};
int[] two = new int[]{3, 4};
List<IntIntPair> list = new ArrayList<>();
IntStream.of(one).forEach(i ->
        IntStream.of(two).mapToObj(j -> PrimitiveTuples.pair(i, j)).forEach(list::add));
System.out.println(list);
// [1:3, 1:4, 2:3, 2:4, 3:3, 3:4]

Unfortunately, I couldn't use flatMap on IntStream as it returns an IntStream . There is no flatMapToObj currently on IntStream , which is what would be needed here. So I used forEach instead.

The IntIntPair and PrimitiveTuples classes I used from Eclipse Collections , as they made it simpler to just output the list as a string. You could use int[] as you have in your solution. The code would look as follows.

List<int[]> list = new ArrayList<>();
IntStream.of(one).forEach(i ->
        IntStream.of(two).mapToObj(j -> new int[]{i, j}).forEach(list::add));

In the 8.1 release of Eclipse Collections (to be released mid-March), there is now a flatCollect method on all primitive containers in the library which can be used to solve this problem. This essentially does what a flatMapToObj method on IntStream should do.

IntList a = IntLists.mutable.with(1, 2, 3);
IntList b = IntLists.mutable.with(3, 4);
List<IntIntPair> result =
        a.flatCollect(
                i -> b.collect(j -> PrimitiveTuples.pair(i, j)),
                Lists.mutable.empty());
System.out.println(result);
// [1:3, 1:4, 2:3, 2:4, 3:3, 3:4]

Update:

As pointed out in the comments by Boris the Spider, the forEach solution would not be thread-safe and would break if the IntStream was parallel . The following solution should work in serial or parallel. I'm glad this was pointed out, because I hadn't thought to do a mapToObj on the IntStream and then followed by a flatMap .

int[] one = new int[]{1, 2, 3};
int[] two = new int[]{3, 4};
List<int[]> list = IntStream.of(one).parallel()
        .mapToObj(i -> IntStream.of(two).mapToObj(j -> new int[]{i, j}))
        .flatMap(e -> e)
        .collect(Collectors.toList());
list.stream().map(e -> "{" + e[0] + "," + e[1] + "}").forEach(System.out::println);

Note: I am a committer for Eclipse Collections .

Answer 4

In this particular case using flatMap you can even bypass an array creation and make your code simpler like this:

list1.stream()
     .flatMap(i -> list2.stream().map(j -> "{" + i+ "," + j + "}"))
     .forEach(System.out::println);

Answer 5

//return pair of numbers
List<List<Integer>> pairs=numbers.stream()
        .flatMap(i -> numbers2.stream()
        .map(j -> Arrays.asList(i,j)))
        .collect(Collectors.toList());

pairs.stream().forEach(System.out::println);

Answer 6

You can generate a 2d list of possible combinations using map and reduce methods:

List<Integer> list1 = Arrays.asList(1, 2, 3);
List<Integer> list2 = Arrays.asList(3, 4);

List<List<Integer>> combinations = Stream.of(list1, list2)
        // represent each list element as a singleton list
        .map(list -> list.stream().map(Collections::singletonList)
                // Stream<List<List<Integer>>>
                .collect(Collectors.toList()))
        // intermediate output
        //[[1], [2], [3]]
        //[[3], [4]]
        .peek(System.out::println)
        // summation of pairs of inner lists
        .reduce((listA, listB) -> listA.stream()
                // combinations of inner lists
                .flatMap(inner1 -> listB.stream()
                        // merge two inner lists into one
                        .map(inner2 -> Stream.of(inner1, inner2)
                                .flatMap(List::stream)
                                .collect(Collectors.toList())))
                // list of combinations
                .collect(Collectors.toList()))
        // otherwise an empty list
        .orElse(Collections.emptyList());

// output
System.out.println(combinations);
// [[1, 3], [1, 4], [2, 3], [2, 4], [3, 3], [3, 4]]

---

^{See also: Generate all combinations from multiple lists}

Answer 7

Of course, you can create a Stream everytime in the Stream#flatMap but in terms of performance, it's very poor.

Instead, I'd advise you to choose for java-16 's Stream#mapMulti and operator on a more declarative way

List<int[]> pairs = list1.stream()
                         .mapMulti((Integer left, Consumer<int[]> consumer) -> 
                             list2.forEach(right -> consumer.accept(new int[]{left, right})))
                         .toList();

Which produces the same result, without the overhead of Stream creation

{1,3}
{1,4}
{2,3}
{2,4}
{3,3}
{3,4}