如何以并发方式在同一个流上执行多种操作？--复制流

正常情况下，一个流在执行一次终端操作之后便结束了。本文通过复制流内数据的方式，曲折的实现了同一个流上执行多次操作。
Demo只是思路，其性能并不一定高效，尤其是数据都在内存中处理时复制的开销很大。但如果流涉及大量I／O，也许性能会有提高。

public class StreamForker<T> {
    private final Stream<T> stream;
    private final Map<Object, Function<Stream<T>, ?>> forks = new HashMap<>();

    public StreamForker(Stream<T> stream) {
        this.stream = stream;
    }

    public StreamForker<T> fork(Object key, Function<Stream<T>, ?> f) {
        forks.put(key, f);
        return this;
    }

    public Results getResults() {
        ForkingStreamConsumer<T> consumer = build();
        try {
            stream.sequential().forEach(consumer);
        } finally {
            consumer.finish();
        }
        return consumer;
    }

    private ForkingStreamConsumer<T> build() {
        List<BlockingQueue<T>> queues = new ArrayList<>();

        Map<Object, Future<?>> actions = forks.entrySet().stream().reduce(new HashMap<Object, Future<?>>(),
                (map, e) -> {
                    map.put(e.getKey(), getOperationResult(queues, e.getValue()));
                    return map;
                }, (m1, m2) -> {
                    m1.putAll(m2);
                    return m1;
                });

        return new ForkingStreamConsumer<>(queues, actions);
    }

    private Future<?> getOperationResult(List<BlockingQueue<T>> queues, Function<Stream<T>, ?> f) {
        BlockingQueue<T> queue = new LinkedBlockingQueue<>();
        queues.add(queue);
        Spliterator<T> spliterator = new BlockingQueueSpliterator<>(queue);
        Stream<T> source = StreamSupport.stream(spliterator, false);
        return CompletableFuture.supplyAsync(() -> f.apply(source));
    }
}

accept方法将原始流中所有的数据添加到各个BlockingQueue内，此处实现了复制

class ForkingStreamConsumer<T> implements Consumer<T>, Results {
    static final Object END_OF_STREAM = new Object();

    private final List<BlockingQueue<T>> queues;
    private final Map<Object, Future<?>> actions;

    public ForkingStreamConsumer(List<BlockingQueue<T>> queues, Map<Object, Future<?>> actions) {
        this.queues = queues;
        this.actions = actions;
    }

    @Override
    public void accept(T t) {
        queues.forEach(q -> q.add(t));
    }

    @SuppressWarnings("unchecked")
    void finish() {
        accept((T) END_OF_STREAM);
    }

    @SuppressWarnings("unchecked")
    @Override
    public <R> R get(Object key) {
        try {
            return ((Future<R>) actions.get(key)).get();
        } catch (Exception e) {
            throw new RuntimeException(e);
        }
    }
}

此处重写了tryAdvance接口，只是简单的从BlockingQueue中取出数据，执行action。业务逻辑中复制流是为了做什么事情，action就是这件事情。ForkingStreamConsumer.END_OF_STREAM是Queue中数据结束的标示

class BlockingQueueSpliterator<T> implements Spliterator<T> {
    private final BlockingQueue<T> q;

    BlockingQueueSpliterator(BlockingQueue<T> q) {
        this.q = q;
    }

    @Override
    public boolean tryAdvance(Consumer<? super T> action) {
        T t;
        while (true) {
            try {
                t = q.take();
                break;
            } catch (InterruptedException e) {
            }
        }

        if (t != ForkingStreamConsumer.END_OF_STREAM) {
            action.accept(t);
            return true;
        }

        return false;
    }

    @Override
    public Spliterator<T> trySplit() {
        return null;
    }

    @Override
    public long estimateSize() {
        return 0;
    }

    @Override
    public int characteristics() {
        return 0;
    }
}