从Java 8流中获取下一个项目

时间:2014-10-27 19:08:51

标签: java java-8 java-stream

我想从Java 8 Stream中检索并删除下一项,而Stream不会被关闭。

Stream<Integer> integerStream = Stream.iterate( 0, x -> new Integer(x + 1) );
Integer zero = integerStream.getNext(); // 0
Integer one  = integerStream.getNext(); // 1
...

这可能吗?

3 个答案:

答案 0 :(得分:14)

是的,有一种方法可以做到这一点,但有一些限制。

Stream<Integer> infiniteStream = Stream.iterate( 0, x -> new Integer(x + 1) );
Iterator<Integer> iter = infiniteStream.iterator();
Integer zero = iter.next();
Integer one  = iter.next();

可替换地,

Stream<Integer> infiniteStream = Stream.iterate( 0, x -> new Integer(x + 1) );
Spliterator<Integer> spliterator = infiniteStream.spliterator();
spliterator.tryAdvance(i -> System.out.println(i)); // zero
spliterator.tryAdvance(i -> System.out.println(i)); // one

如果有Stream,则可以从中获取IteratorSpliterator,或查询它是否为并行流等。这些是在BaseStream接口上定义,是Stream的超接口,这使得它们容易错过。

在这种情况下,我们知道流是无限的,因此无需调用Iterator的hasNext()方法或检查Spliterator tryAdvance()的返回值

限制是iterator()的{​​{1}}和spliterator()方法都是终端操作,这意味着在他们被调用之后,返回Iterator或Spliterator可以独占访问Stream表示的值。不允许对流进行进一步操作(例如Streamfilter等),并且map将会遇到这些操作。

如果您想剥离前几个元素,然后恢复流处理,可以将分裂器重新转换为如下所示的流:

IllegalStateException

这可能对某些事情很好,但我不确定我是否会推荐这种技术。我认为它会在生成下一个元素的路径中添加一些额外的对象,从而增加额外的方法调用。

编辑评论(与您的问题无关):

  • 不要使用Stream<Integer> stream2 = StreamSupport.stream(spliterator, false); 。而是使用new Integer(val),如果它可用,它将重复使用盒装整数,对于-128到127范围内的值通常是正确的。
  • 您可以使用Integer.valueOf(val)代替IntStream,这样可以完全避免装箱费用。它没有完整的流操作,但它有Stream<Integer>,它采用一个对原始iterate()值进行操作的函数。

答案 1 :(得分:1)

基于Stuart's answerIterator-to-Stream conversion,我提出了以下快速而肮脏的包装类。它未经过测试,并且不是线程安全的,但它为我提供了我目前所需的内容 - 删除和使用单个项目,同时保留流&#34;打开& #34 ;.

PeelingStream<T>提供了一种方法T getNext(),可以屏蔽someWrappedStream.iterator() 终端流操作语义:

public class PeelingStream<T> implements Stream<T> {

    private Stream<T> wrapped;

    public PeelingStream(Stream<T> toBeWrapped) {
        this.wrapped = toBeWrapped;
    }

    public T getNext() {
        Iterator<T> iterator = wrapped.iterator();
        T next = iterator.next();
        Iterable<T> remainingIterable = () -> iterator;
        wrapped = StreamSupport.stream(remainingIterable.spliterator(),
                false);

        return next;
    }

    ///////////////////// from here, only plain delegate methods

    public Iterator<T> iterator() {
        return wrapped.iterator();
    }

    public Spliterator<T> spliterator() {
        return wrapped.spliterator();
    }

    public boolean isParallel() {
        return wrapped.isParallel();
    }

    public Stream<T> sequential() {
        return wrapped.sequential();
    }

    public Stream<T> parallel() {
        return wrapped.parallel();
    }

    public Stream<T> unordered() {
        return wrapped.unordered();
    }

    public Stream<T> onClose(Runnable closeHandler) {
        return wrapped.onClose(closeHandler);

    }

    public void close() {
        wrapped.close();
    }

    public Stream<T> filter(Predicate<? super T> predicate) {
        return wrapped.filter(predicate);
    }

    public <R> Stream<R> map(Function<? super T, ? extends R> mapper) {
        return wrapped.map(mapper);
    }

    public IntStream mapToInt(ToIntFunction<? super T> mapper) {
        return wrapped.mapToInt(mapper);
    }

    public LongStream mapToLong(ToLongFunction<? super T> mapper) {
        return wrapped.mapToLong(mapper);
    }

    public DoubleStream mapToDouble(ToDoubleFunction<? super T> mapper) {
        return wrapped.mapToDouble(mapper);
    }

    public <R> Stream<R> flatMap(
            Function<? super T, ? extends Stream<? extends R>> mapper) {
        return wrapped.flatMap(mapper);
    }

    public IntStream flatMapToInt(
            Function<? super T, ? extends IntStream> mapper) {
        return wrapped.flatMapToInt(mapper);
    }

    public LongStream flatMapToLong(
            Function<? super T, ? extends LongStream> mapper) {
        return wrapped.flatMapToLong(mapper);
    }

    public DoubleStream flatMapToDouble(
            Function<? super T, ? extends DoubleStream> mapper) {
        return wrapped.flatMapToDouble(mapper);
    }

    public Stream<T> distinct() {
        return wrapped.distinct();
    }

    public Stream<T> sorted() {
        return wrapped.sorted();
    }

    public Stream<T> sorted(Comparator<? super T> comparator) {
        return wrapped.sorted(comparator);
    }

    public Stream<T> peek(Consumer<? super T> action) {
        return wrapped.peek(action);
    }

    public Stream<T> limit(long maxSize) {
        return wrapped.limit(maxSize);
    }

    public Stream<T> skip(long n) {
        return wrapped.skip(n);
    }

    public void forEach(Consumer<? super T> action) {
        wrapped.forEach(action);
    }

    public void forEachOrdered(Consumer<? super T> action) {
        wrapped.forEachOrdered(action);
    }

    public Object[] toArray() {
        return wrapped.toArray();
    }

    public <A> A[] toArray(IntFunction<A[]> generator) {
        return wrapped.toArray(generator);
    }

    public T reduce(T identity, BinaryOperator<T> accumulator) {
        return wrapped.reduce(identity, accumulator);
    }

    public Optional<T> reduce(BinaryOperator<T> accumulator) {
        return wrapped.reduce(accumulator);
    }

    public <U> U reduce(U identity,
            BiFunction<U, ? super T, U> accumulator,
            BinaryOperator<U> combiner) {
        return wrapped.reduce(identity, accumulator, combiner);
    }

    public <R> R collect(Supplier<R> supplier,
            BiConsumer<R, ? super T> accumulator, BiConsumer<R, R> combiner) {
        return wrapped.collect(supplier, accumulator, combiner);
    }

    public <R, A> R collect(Collector<? super T, A, R> collector) {
        return wrapped.collect(collector);
    }

    public Optional<T> min(Comparator<? super T> comparator) {
        return wrapped.min(comparator);
    }

    public Optional<T> max(Comparator<? super T> comparator) {
        return wrapped.max(comparator);
    }

    public long count() {
        return wrapped.count();
    }

    public boolean anyMatch(Predicate<? super T> predicate) {
        return wrapped.anyMatch(predicate);
    }

    public boolean allMatch(Predicate<? super T> predicate) {
        return wrapped.allMatch(predicate);
    }

    public boolean noneMatch(Predicate<? super T> predicate) {
        return wrapped.noneMatch(predicate);
    }

    public Optional<T> findFirst() {
        return wrapped.findFirst();
    }

    public Optional<T> findAny() {
        return wrapped.findAny();
    }

}

小测试:

@Test
public void testPeelingOffItemsFromStream() {

    Stream<Integer> infiniteStream = Stream.iterate(0, x -> x + 1);

    PeelingStream<Integer> peelingInfiniteStream = new PeelingStream<>(infiniteStream);

    Integer one = peelingInfiniteStream.getNext();
    assertThat(one, equalTo(0));

    Integer two = peelingInfiniteStream.getNext();
    assertThat(two, equalTo(1));

    Stream<Integer> limitedStream = peelingInfiniteStream.limit(3); // 2 3 4
    int sumOf234 = limitedStream.mapToInt(x -> x.intValue()).sum();
    assertThat(sumOf234, equalTo(2 + 3 + 4));

}

答案 2 :(得分:0)

我做了以下事情。原始流确实关闭了,但创建了一个与旧流完全相同的新流。

您将需要来自番石榴的 com.google.common.collect.Iterators。

import static com.google.common.collect.Iterators.concat;
import static com.google.common.collect.Iterators.singletonIterator;
import static java.util.Spliterators.spliteratorUnknownSize;
import static java.util.stream.StreamSupport.stream;

private <T> Stream<T> peekFirst(Stream<T> originalStream){
    //This closes the original Stream
    Iterator<T> originalIterator = originalStream.iterator();
    if (!originalIterator.hasNext()) {
        return Stream.of();
    }
    T firstElement = originalIterator.next();
    doSomethingWithFirstElement(firstElement);
    Iterator<T> newIterator = concat(
      singletonIterator(firstElement),
      originalIterator);
    return stream(
      spliteratorUnknownSize(newIterator, 0), 
        originalStream.isParallel());
}