CompletableFuture 组合式异步编程

​​CompletableFuture​​​是jdk1.8引入的一个新特性。 它主要是为了解决多个​​Future​​结果之间的依赖关系。比如：

• 将两个异步计算合并为一个——这两个异步计算之间相互独立，同时第二个又依赖于第
  一个的结果。
• 等待​​Future​​集合中的所有任务都完成。
• 仅等待​​Future​​​集合中最快结束的任务完成（有可能因为它们试图通过不同的方式计算同
  一个值），并返回它的结果。
• 通过编程方式完成一个​​Future​​任务的执行（即以手工设定异步操作结果的方式）。
• 应对Future的完成事件（即当​​Future​​​的完成事件发生时会收到通知，并能使用​​Future​​ 计算的结果进行下一步的操作，不只是简单地阻塞等待操作的结果）。

runAsync 和 supplyAsync方法

​​CompletableFuture​​提供了四个静态方法来创建一个异步操作。

public static CompletableFuture<Void> runAsync(Runnable runnable)
public static CompletableFuture<Void> runAsync(Runnable runnable, Executor executor)
public static <U> CompletableFuture<U> supplyAsync(Supplier<U> supplier)
public static <U> CompletableFuture<U> supplyAsync(Supplier<U> supplier, Executor executor)

• ​​runAsync​​方法不支持返回值
• ​​supplyAsync​​可以支持返回值

如果​​executor​​​参数没有设置值，那么会使用​​ForkJoinPool.commonPool​​默认线程池执行任务。

示例

@Test
public void testRunAsync() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    CompletableFuture.runAsync(() -> System.out.println(Thread.currentThread().getName() +  " 执行异步任务 runAsync"), executor);
    String result = CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync");
        return "有结果";
    }, executor).get();
    System.out.println(result);
}

pool-1-thread-1 执行异步任务 runAsync
pool-1-thread-2 执行异步任务 supplyAsync
有结果

计算结果完成时的回调方法

当​​CompletableFuture​​的计算结果完成，或者抛出异常的时候，可以执行特定的Action。主要是下面的方法：

public CompletableFuture<T> whenComplete(BiConsumer<? super T,? super Throwable> action)
public CompletableFuture<T> whenCompleteAsync(BiConsumer<? super T,? super Throwable> action)
public CompletableFuture<T> whenCompleteAsync(BiConsumer<? super T,? super Throwable> action, Executor executor)
public CompletableFuture<T> exceptionally(Function<Throwable,? extends T> fn)

可以看到Action的类型是​​BiConsumer<? super T,? super Throwable>​​​它可以处理正常的计算结果，或者异常情况。
​​​whenComplete​​​和 ​​whenCompleteAsync​​的区别：

• ​​whenComplete​​：是执行当前任务的线程执行继续执行 whenComplete 的任务。
• ​​whenCompleteAsync​​：是执行把 whenCompleteAsync 这个任务继续提交给线程池来进行执行。

示例

@Test
public void testWhenComplete() {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync");
        return "有结果";
    }, executor).whenComplete((s, throwable) -> {
        System.out.println(Thread.currentThread().getName() + " 执行了 whenComplete");
        if (!StringUtils.isEmpty(s)) {
            System.out.println(Thread.currentThread().getName() + " 真的有结果诶！ 结果是:" + s);
        }
    }).whenCompleteAsync((s, throwable) -> {
        System.out.println(Thread.currentThread().getName() + " 执行了 whenCompleteAsync");
        System.out.println(Thread.currentThread().getName() + " 打印结果值 :" + s);
    }, executor).exceptionally(throwable -> {
        System.out.println(Thread.currentThread().getName() + " 执行了 exceptionally");
        System.out.println(Thread.currentThread().getName() + " 异常了 :" + throwable.getMessage());
        return "异常了";
    });
}

pool-1-thread-1 执行异步任务 supplyAsync
pool-1-thread-1 执行了 whenComplete
pool-1-thread-1 真的有结果诶！ 结果是:有结果
pool-1-thread-2 执行了 whenCompleteAsync
pool-1-thread-2 打印结果值 :有结果

thenApply 方法

当一个线程依赖另一个线程时，可以使用 thenApply 方法来把这两个线程串行化。

public <U> CompletableFuture<U> thenApply(Function<? super T,? extends U> fn)
public <U> CompletableFuture<U> thenApplyAsync(Function<? super T,? extends U> fn)
public <U> CompletableFuture<U> thenApplyAsync(Function<? super T,? extends U> fn, Executor executor)

​​Function<? super T,? extends U>​​

• T：上一个任务返回结果的类型
• U：当前任务的返回值类型

​​thenApply​​​和​​thenApplyAsync​​的区别：

• ​​thenApply​​：使用当前线程来执行任务
• ​​thenApplyAsync​​​：如果设置了​​executor​​​则使用设置的线程池执行任务，如果没有设置则使用​​ForkJoinPool.commonPool​​线程池执行

示例

@Test
public void testThenApply() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    Integer integer = CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync");
        return "有结果";
    }, executor).thenApplyAsync(s -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 thenApply");
        if (!StringUtils.isEmpty(s)) {
            return s.length();
        }
        return 0;
    }, executor).get();
    System.out.println(Thread.currentThread().getName() + " " + integer);
}

pool-1-thread-1 执行异步任务 supplyAsync
pool-1-thread-2 执行异步任务 thenApply
main 3

handle 方法

​​handle​​​是执行任务完成时对结果的处理。
​​​handle​​​方法和​​thenApply​​​方法处理方式基本一样。不同的是 ​​handle​​​是在任务完成后再执行，还可以处理异常的任务。​​thenApply​​​只可以执行正常的任务，任务出现异常则不执行 ​​thenApply​​方法。

public <U> CompletionStage<U> handle(BiFunction<? super T, Throwable, ? extends U> fn);
public <U> CompletionStage<U> handleAsync(BiFunction<? super T, Throwable, ? extends U> fn);
public <U> CompletionStage<U> handleAsync(BiFunction<? super T, Throwable, ? extends U> fn,Executor executor);

示例

@Test
public void testHandle() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    Integer integer = CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync");
        return 10/0;
    }, executor).handleAsync((s, throwable) -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 handleAsync");

        if (Objects.nonNull(throwable)) {
            System.out.println(Thread.currentThread().getName() + " 异常信息 " + throwable.getMessage());
            return 1;
        }

        return s;
    }, executor).get();
    System.out.println(Thread.currentThread().getName() + " " + integer);
}

pool-1-thread-1 执行异步任务 supplyAsync
pool-1-thread-2 执行异步任务 handleAsync
pool-1-thread-2 异常信息 java.lang.ArithmeticException: / by zero
main 1

从示例中可以看出，在​​handle​​​中可以根据任务是否有异常来进行做相应的后续处理操作。而 ​​thenApply​​​方法，如果上个任务出现错误，则不会执行 ​​thenApply​​方法。

thenAccept 消费处理结果

接收任务的处理结果，并消费处理，无返回结果。

public CompletionStage<Void> thenAccept(Consumer<? super T> action);
public CompletionStage<Void> thenAcceptAsync(Consumer<? super T> action);
public CompletionStage<Void> thenAcceptAsync(Consumer<? super T> action,Executor executor);

示例

@Test
@Test
public void testThenAccept() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync");
        return 10;
    }, executor).thenAcceptAsync(s -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 thenAccept");
        System.out.println(Thread.currentThread().getName() + " " + s);
    }, executor).get();
}

pool-1-thread-1 执行异步任务 supplyAsync
pool-1-thread-2 执行异步任务 thenAccept
pool-1-thread-2 10

从示例代码中可以看出，该方法只是消费执行完成的任务，并可以根据上面的任务返回的结果进行处理。如果第一个任务发生异常，那么​​thenAccept​​方法不会被执行。

thenRun 方法

跟 ​​thenAccept​​​方法不一样的是，​​thenRun​​​不关心任务的处理结果。只要上面的任务执行完成，就开始执行 ​​thenRun​​里面的任务。

public CompletionStage<Void> thenRun(Runnable action);
public CompletionStage<Void> thenRunAsync(Runnable action);
public CompletionStage<Void> thenRunAsync(Runnable action,Executor executor);

示例

@Test
public void testThenRun() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync");
        return 10;
    }, executor).thenRunAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 thenRun");
    }, executor).get();
}

pool-1-thread-1 执行异步任务 supplyAsync
pool-1-thread-2 执行异步任务 thenRun

thenCombine 合并任务

​​thenCombine​​​会把 两个​​CompletionStage​​​的任务都执行完成后，把两个任务的结果一块交给​​thenCombine​​来处理。

public <U,V> CompletionStage<V> thenCombine(CompletionStage<? extends U> other,BiFunction<? super T,? super U,? extends V> fn);
public <U,V> CompletionStage<V> thenCombineAsync(CompletionStage<? extends U> other,BiFunction<? super T,? super U,? extends V> fn);
public <U,V> CompletionStage<V> thenCombineAsync(CompletionStage<? extends U> other,BiFunction<? super T,? super U,? extends V> fn,Executor executor);

示例

@Test
public void testThenCombine() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    Integer result = CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync1");
        return 10;
    }, executor).thenCombine(CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync2");
        return 15;
    }, executor), (result1, result2) -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 thenCombine");
        return result1 + result2;
    }).get();
    System.out.println(Thread.currentThread().getName() + "  " + result);

    // 分解一下好看些
    CompletableFuture<Integer> future1 = CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync1");
        return 10;
    }, executor);
    CompletableFuture<Integer> future2 = CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync2");
        return 15;
    }, executor);

    result = future1.thenCombine(future2, (result1, result2) -> {
                System.out.println(Thread.currentThread().getName() + " 执行异步任务 thenCombine");
                return result1 + result2;
            }).get();
    System.out.println(Thread.currentThread().getName() + "  " + result);
}

pool-1-thread-1 执行异步任务 supplyAsync1
pool-1-thread-2 执行异步任务 supplyAsync2
pool-1-thread-2 执行异步任务 thenCombine
main  25
pool-1-thread-3 执行异步任务 supplyAsync1
pool-1-thread-4 执行异步任务 supplyAsync2
main 执行异步任务 thenCombine
main  25

thenAcceptBoth

当两个​​CompletionStage​​​都执行完成后，把结果一块交给​​thenAcceptBoth​​来进行消耗。

public <U> CompletionStage<Void> thenAcceptBoth(CompletionStage<? extends U> other,BiConsumer<? super T, ? super U> action);
public <U> CompletionStage<Void> thenAcceptBothAsync(CompletionStage<? extends U> other,BiConsumer<? super T, ? super U> action);
public <U> CompletionStage<Void> thenAcceptBothAsync(CompletionStage<? extends U> other,BiConsumer<? super T, ? super U> action,     Executor executor);

示例

@Test
public void testThenAcceptBoth() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);

    // 分解一下好看些
    CompletableFuture<Integer> future1 = CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync1");
        return 10;
    }, executor);
    CompletableFuture<Integer> future2 = CompletableFuture.supplyAsync(() -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync2");
        return 15;
    }, executor);

    future1.thenAcceptBoth(future2, (result1, result2) -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 thenCombine");
        System.out.println(Thread.currentThread().getName() + " " + (result1 + result2));
    }).get();
}

pool-1-thread-1 执行异步任务 supplyAsync1
pool-1-thread-2 执行异步任务 supplyAsync2
pool-1-thread-2 执行异步任务 thenAcceptBoth
pool-1-thread-2 25

applyToEither 方法

两个​​CompletionStage​​​，谁执行返回的结果快，我就用那个​​CompletionStage​​的结果进行下一步的转化操作。

public <U> CompletionStage<U> applyToEither(CompletionStage<? extends T> other,Function<? super T, U> fn);
public <U> CompletionStage<U> applyToEitherAsync(CompletionStage<? extends T> other,Function<? super T, U> fn);
public <U> CompletionStage<U> applyToEitherAsync(CompletionStage<? extends T> other,Function<? super T, U> fn,Executor executor);

示例

@Test
public void testApplyToEither() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    // 分解一下好看些
    CompletableFuture<Integer> future1 = CompletableFuture.supplyAsync(() -> {
        int probe  = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync1   " + probe);
        CompletableFutureTest.sleep(probe);
        return probe;
    }, executor);

    CompletableFuture<Integer> future2 = CompletableFuture.supplyAsync(() -> {
        int probe  = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync2   " + probe);
        CompletableFutureTest.sleep(probe);
        return probe;
    }, executor);

    String result = future1.applyToEither(future2, integer -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 applyToEither");
        return integer + "";
    }).get();
    System.out.println(Thread.currentThread().getName() + " 执行异步任务 " + result);
}

pool-1-thread-1 执行异步任务 supplyAsync1   301
pool-1-thread-2 执行异步任务 supplyAsync2   422
pool-1-thread-1 执行异步任务 applyToEither
main 执行异步任务 301

acceptEither 方法

两个​​CompletionStage​​​，谁执行返回的结果快，我就用那个​​CompletionStage​​的结果进行下一步的消耗操作。

public CompletionStage<Void> acceptEither(CompletionStage<? extends T> other,Consumer<? super T> action);
public CompletionStage<Void> acceptEitherAsync(CompletionStage<? extends T> other,Consumer<? super T> action);
public CompletionStage<Void> acceptEitherAsync(CompletionStage<? extends T> other,Consumer<? super T> action,Executor executor);

示例

@Test
public void testAcceptEither() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    // 分解一下好看些
    CompletableFuture<Integer> future1 = CompletableFuture.supplyAsync(() -> {
        int probe = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync1   " + probe);
        CompletableFutureTest.sleep(probe);
        return probe;
    }, executor);
    CompletableFuture<Integer> future2 = CompletableFuture.supplyAsync(() -> {
        int probe = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync2   " + probe);
        CompletableFutureTest.sleep(probe);
        return probe;
    }, executor);

    future1.acceptEither(future2, integer -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 acceptEither" + integer);
    }).get();
}

pool-1-thread-1 执行异步任务 supplyAsync1   507
pool-1-thread-2 执行异步任务 supplyAsync2   167
pool-1-thread-2 执行异步任务 acceptEither   167

runAfterEither 方法

两个​​CompletionStage​​​，任何一个完成了都会执行下一步的操作（​​Runnable​​）

public CompletionStage<Void> runAfterEither(CompletionStage<?> other,Runnable action);
public CompletionStage<Void> runAfterEitherAsync(CompletionStage<?> other,Runnable action);
public CompletionStage<Void> runAfterEitherAsync(CompletionStage<?> other,Runnable action,Executor executor);

示例

@Test
public void testRunAfterEither() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    long start = System.currentTimeMillis();
    // 分解一下好看些
    CompletableFuture<Integer> future1 = CompletableFuture.supplyAsync(() -> {
        int probe = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync1   " + probe);
        CompletableFutureTest.sleep(probe);
        return probe;
    }, executor);
    CompletableFuture<Integer> future2 = CompletableFuture.supplyAsync(() -> {
        int probe = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync2   " + probe);
        CompletableFutureTest.sleep(probe);
        return probe;
    }, executor);

    future1.runAfterEither(future2, () -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 runAfterEither  ");
    }).get();

    System.out.println(Thread.currentThread().getName() + " 任务结束 耗时：" + (System.currentTimeMillis() - start));
}

pool-1-thread-1 执行异步任务 supplyAsync1   23
pool-1-thread-2 执行异步任务 supplyAsync2   704
pool-1-thread-1 执行异步任务 runAfterEither  
main 任务结束 耗时：32

runAfterBoth 方法

两个​​CompletionStage​​，都完成了计算才会执行下一步的操作（Runnable）

public CompletionStage<Void> runAfterBoth(CompletionStage<?> other,Runnable action);
public CompletionStage<Void> runAfterBothAsync(CompletionStage<?> other,Runnable action);
public CompletionStage<Void> runAfterBothAsync(CompletionStage<?> other,Runnable action,Executor executor);

示例

@Test
public void testRunAfterBoth () throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    long start = System.currentTimeMillis();
    // 分解一下好看些
    CompletableFuture<Integer> future1 = CompletableFuture.supplyAsync(() -> {
        int probe = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync1   " + probe);
        CompletableFutureTest.sleep(probe);
        return probe;
    }, executor);
    CompletableFuture<Integer> future2 = CompletableFuture.supplyAsync(() -> {
        int probe = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync2   " + probe);
        CompletableFutureTest.sleep(probe);
        return probe;
    }, executor);

    future1.runAfterBoth(future2, () -> {
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 runAfterBoth   ");
    }).get();

    System.out.println(Thread.currentThread().getName() + " 任务结束 耗时：" + (System.currentTimeMillis() - start));
}

pool-1-thread-1 执行异步任务 supplyAsync1   576
pool-1-thread-2 执行异步任务 supplyAsync2   941
pool-1-thread-2 执行异步任务 runAfterBoth   
main 任务结束 耗时：959

thenCompose 方法

​​thenCompose​​​方法允许你对两个​​CompletionStage​​进行流水线操作，第一个操作完成时，将其结果作为参数传递给第二个操作。

public <U> CompletableFuture<U> thenCompose(Function<? super T, ? extends CompletionStage<U>> fn);
public <U> CompletableFuture<U> thenComposeAsync(Function<? super T, ? extends CompletionStage<U>> fn) ;
public <U> CompletableFuture<U> thenComposeAsync(Function<? super T, ? extends CompletionStage<U>> fn, Executor executor) ;

示例

@Test
public void testThenCompose() throws ExecutionException, InterruptedException {
    ExecutorService executor = Executors.newFixedThreadPool(10);
    long start = System.currentTimeMillis();
    // 分解一下好看些
    CompletableFuture<Integer> future1 = CompletableFuture.supplyAsync(() -> {
        int probe = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync1   " + probe);
        CompletableFutureTest.sleep(probe);
        return probe;
    }, executor);


    String result = future1.thenCompose(integer -> CompletableFuture.supplyAsync(() -> {
        int probe = ThreadLocalRandom.current().nextInt(1000);
        System.out.println(Thread.currentThread().getName() + " 执行异步任务 supplyAsync1   " + probe);
        CompletableFutureTest.sleep(probe);
        return integer + "  " + probe;
    }, executor)).get();

    System.out.println(Thread.currentThread().getName() + "  " + result);
    System.out.println(Thread.currentThread().getName() + " 任务结束 耗时：" + (System.currentTimeMillis() - start));
}

pool-1-thread-1 执行异步任务 supplyAsync1   460
pool-1-thread-2 执行异步任务 supplyAsync1   320
main  460  320
main 任务结束 耗时：788

总结

• 如果没有指定线程池​​CompletableFuture​​​会使用​​ForkJoinPool.commonPool()​​作为它的线程池执行异步代码，如果指定线程池，则使用指定的线程池运行。

参考

《Java 8实战》
​​​ https://www.jianshu.com/p/6bac52527ca4​​

源码

​​https://github.com/wyh-spring-ecosystem-student/spring-boot-student/tree/releases​​

spring-boot-student-concurrent 工程

layering-cache

为监控而生的多级缓存框架 layering-cache这是我开源的一个多级缓存框架的实现，如果有兴趣可以看一下