Java多线程的Callable, Future, FutureCallback, CompletableFuture
Callable可以看成是一个增强版的Runnable, 带返回结果, 需要通过Future或者FutureTask来提交任务或运行线程, 然后通过Future/FutureTask的get方法得到返回结果.
Callable在子线程中运行, 在主线程中异步得到执行结果(get()方法是阻塞的), 或者检查是否已取消, 是否已完成(检查取消和完成的方法是非阻塞的)
通过Thread子线程启动
这种方式, 需要创建一个FutureTask对象, 再用这个FutureTask对象创建一个Thread来运行. 后续操作都通过FutureTask进行.
public class DemoCallableFuture {
public static void main(String[] args) {
FutureTask<String> task = new FutureTask<>(()->{
System.out.println("task start");
Thread.sleep(1000);
System.out.println("task done");
return "task get";
});
new Thread(task).start();
FutureTask<String> task2 = new FutureTask<>(()->{
System.out.println("task2 start");
Thread.sleep(1000);
System.out.println("task2 done");
return "task2 get";
});
new Thread(task2).start();
if (task.isCancelled()) {
System.out.println("task cancelled yes");
} else {
System.out.println("task cancelled no");
}
if (task.isDone()) {
System.out.println("task done yes");
} else {
System.out.println("task done no");
}
try {
System.out.println(task.get());
} catch (InterruptedException|ExecutionException e) {
e.printStackTrace();
}
if (task2.isCancelled()) {
System.out.println("task2 cancelled yes");
} else {
System.out.println("task2 cancelled no");
}
if (task2.isDone()) {
System.out.println("task2 done yes");
} else {
System.out.println("task2 done no");
}
try {
System.out.println(task2.get());
} catch (InterruptedException|ExecutionException e) {
e.printStackTrace();
}
}
}
运行结果
task start task cancelled no task done no task2 start task2 done task done task get task2 cancelled no task2 done yes task2 get
通过ExecutorCompletionService
ExecutorCompletionService是CompletionService在JUC里唯一的实现类, 这个类实现了非排序异步获取任务结果的功能, 通过ExecutorCompletionService.submit() 可以添加Callable并返回Future, 而通过多次调用 ExecutorCompletionService.take().get(), 可以按任务完成的顺序依次取回结果. 避免了使用Future.get()时, 无法根据结果返回次序获取结果的问题.
通过CompletableFuture
CompletableFuture.supplyAsync()可以输入一个Supplier<U>作为参数, 这个参数和Callable<U>的作用是相似的. 而后可以用thenApply加上结果处理方法, 适合全异步的处理, 例如
public void asyncGet(String url, Charset charset, Function<Response, Response> function) {
CompletableFuture.supplyAsync(() -> get(url, charset)).thenApply(function);
}
通过ExecutorService线程池启动
这种方式, 通过线程池submit一个Callable对象, 就会得到一个Future对象, 根据这个Future对象做后续操作
public class DemoCallableFuture2 {
public static void main(String[] args) {
ExecutorService service = Executors.newFixedThreadPool(4);
Future<String> future = service.submit(()->{
System.out.println("task start");
Thread.sleep(1000);
System.out.println("task done");
return "task get";
});
if (future.isCancelled()) {
System.out.println("task cancelled yes");
} else {
System.out.println("task cancelled no");
}
if (future.isDone()) {
System.out.println("task done yes");
} else {
System.out.println("task done no");
}
try {
Thread.sleep(1000);
} catch (InterruptedException e) {
e.printStackTrace();
}
if (future.isCancelled()) {
System.out.println("task cancelled yes");
} else {
System.out.println("task cancelled no");
}
if (future.isDone()) {
System.out.println("task done yes");
} else {
System.out.println("task done no");
}
try {
System.out.println(future.get());
} catch (InterruptedException|ExecutionException e) {
e.printStackTrace();
}
}
}
运行结果
task cancelled no task done no task start task cancelled no task done no task done task get
.
FutureCallback
FutureCallback是Google Guava中的一个类, 解决的是Future中get阻塞的问题, 让全过程异步. 需要使用ListeningExecutorService的线程池提交.
代码例子
public class DemoFutureCallback {
public static void main(String[] args) {
ListeningExecutorService service = MoreExecutors.listeningDecorator(Executors.newFixedThreadPool(10));
for (int i = 0; i < 5; i++) {
int j = i;
ListenableFuture<String> future = service.submit(()->{
System.out.println("task start");
Thread.sleep(1000);
System.out.println("task done");
return "task return " + j;
});
Futures.addCallback(future, new FutureCallback<String>() {
@Override
public void onSuccess(String s) {
System.out.println("callback success: " + s);
}
@Override
public void onFailure(Throwable throwable) {
throwable.printStackTrace();
}
});
}
System.out.println("thread ongoing");
service.shutdown();
}
}
运行结果
task start task start task start thread ongoing task start task start task done callback success: task return 0 task done callback success: task return 1 task done callback success: task return 3 task done callback success: task return 4 task done callback success: task return 2 Process finished with exit code 0
