下面CompletableFuture例子中join的调用是否阻塞进程
我试图理解 CompletableFutures 和返回已完成期货的调用链，我创建了下面的示例，它模拟了对数据库的两次调用。
第一个方法应该是用 userIds 列表给出一个可完成的未来，然后我需要调用另一个方法提供 userId 来获取用户（在本例中是一个字符串）。
总结一下：
1. 获取 ID
2. 获取与这些 ID 对应的用户列表。
我创建了简单的方法来模拟休眠线程的响应。请检查下面的代码
public class PipelineOfTasksExample {
private Map<Long, String> db = new HashMap<>();
PipelineOfTasksExample() {
db.put(1L, "user1");
db.put(2L, "user2");
db.put(3L, "user3");
db.put(4L, "user4");
}
private CompletableFuture<List<Long>> returnUserIdsFromDb() {
try {
Thread.sleep(500);
} catch (InterruptedException e) {
e.printStackTrace();
}
System.out.println("building the list of Ids" + " - thread: " + Thread.currentThread().getName());
return CompletableFuture.supplyAsync(() -> Arrays.asList(1L, 2L, 3L, 4L));
}
private CompletableFuture<String> fetchById(Long id) {
CompletableFuture<String> cfId = CompletableFuture.supplyAsync(() -> db.get(id));
try {
Thread.sleep(500);
} catch (InterruptedException e) {
e.printStackTrace();
}
System.out.println("fetching id: " + id + " -> " + db.get(id) + " thread: " + Thread.currentThread().getName());
return cfId;
}
public static void main(String[] args) {
PipelineOfTasksExample example = new PipelineOfTasksExample();
CompletableFuture<List<String>> result = example.returnUserIdsFromDb()
.thenCompose(listOfIds ->
CompletableFuture.supplyAsync(
() -> listOfIds.parallelStream()
.map(id -> example.fetchById(id).join())
.collect(Collectors.toList()
)
)
);
System.out.println(result.join());
}
}
我的问题是，join call ( example.fetchById(id).join()) 是否破坏了进程的非阻塞性质。如果答案是肯定的，我该如何解决这个问题？
BIG阳1回答
-
桃花长相依
你的例子有点奇怪，因为你returnUserIdsFromDb()在任何操作甚至开始之前就减慢了主线程的速度，同样，fetchById减慢了调用者而不是异步操作的速度，这违背了异步操作的全部目的。此外，.thenCompose(listOfIds -> CompletableFuture.supplyAsync(() -> …))您可以简单地使用.thenApplyAsync(listOfIds -> …).所以一个更好的例子可能是public class PipelineOfTasksExample { private final Map<Long, String> db = LongStream.rangeClosed(1, 4).boxed() .collect(Collectors.toMap(id -> id, id -> "user"+id)); PipelineOfTasksExample() {} private static <T> T slowDown(String op, T result) { LockSupport.parkNanos(TimeUnit.MILLISECONDS.toNanos(500)); System.out.println(op + " -> " + result + " thread: " + Thread.currentThread().getName()+ ", " + POOL.getPoolSize() + " threads"); return result; } private CompletableFuture<List<Long>> returnUserIdsFromDb() { System.out.println("trigger building the list of Ids - thread: " + Thread.currentThread().getName()); return CompletableFuture.supplyAsync( () -> slowDown("building the list of Ids", Arrays.asList(1L, 2L, 3L, 4L)), POOL); } private CompletableFuture<String> fetchById(Long id) { System.out.println("trigger fetching id: " + id + " thread: " + Thread.currentThread().getName()); return CompletableFuture.supplyAsync( () -> slowDown("fetching id: " + id , db.get(id)), POOL); } static ForkJoinPool POOL = new ForkJoinPool(2); public static void main(String[] args) { PipelineOfTasksExample example = new PipelineOfTasksExample(); CompletableFuture<List<String>> result = example.returnUserIdsFromDb() .thenApplyAsync(listOfIds -> listOfIds.parallelStream() .map(id -> example.fetchById(id).join()) .collect(Collectors.toList() ), POOL ); System.out.println(result.join()); }}打印出类似的东西trigger building the list of Ids - thread: mainbuilding the list of Ids -> [1, 2, 3, 4] thread: ForkJoinPool-1-worker-1, 1 threadstrigger fetching id: 2 thread: ForkJoinPool-1-worker-0trigger fetching id: 3 thread: ForkJoinPool-1-worker-1trigger fetching id: 4 thread: ForkJoinPool-1-worker-2fetching id: 4 -> user4 thread: ForkJoinPool-1-worker-3, 4 threadsfetching id: 2 -> user2 thread: ForkJoinPool-1-worker-3, 4 threadsfetching id: 3 -> user3 thread: ForkJoinPool-1-worker-2, 4 threadstrigger fetching id: 1 thread: ForkJoinPool-1-worker-3fetching id: 1 -> user1 thread: ForkJoinPool-1-worker-2, 4 threads[user1, user2, user3, user4]乍一看，这可能是一个惊人的线程数。答案是join()可能会阻塞线程，但是如果这种情况发生在Fork/Join池的工作线程内部，这种情况会被检测到并启动一个新的补偿线程，以确保配置的目标并行度。作为一种特殊情况，当使用默认的 Fork/Join 池时，实现可能会在方法内选择新的待处理任务join()，以确保同一线程内的进度。所以代码总是会取得进展，join()偶尔调用也没有错，如果替代方案要复杂得多，但如果过度使用，就会有资源消耗过多的危险。毕竟，之所以要使用线程池，就是为了限制线程的数量。另一种方法是尽可能使用链式依赖操作。public class PipelineOfTasksExample { private final Map<Long, String> db = LongStream.rangeClosed(1, 4).boxed() .collect(Collectors.toMap(id -> id, id -> "user"+id)); PipelineOfTasksExample() {} private static <T> T slowDown(String op, T result) { LockSupport.parkNanos(TimeUnit.MILLISECONDS.toNanos(500)); System.out.println(op + " -> " + result + " thread: " + Thread.currentThread().getName()+ ", " + POOL.getPoolSize() + " threads"); return result; } private CompletableFuture<List<Long>> returnUserIdsFromDb() { System.out.println("trigger building the list of Ids - thread: " + Thread.currentThread().getName()); return CompletableFuture.supplyAsync( () -> slowDown("building the list of Ids", Arrays.asList(1L, 2L, 3L, 4L)), POOL); } private CompletableFuture<String> fetchById(Long id) { System.out.println("trigger fetching id: " + id + " thread: " + Thread.currentThread().getName()); return CompletableFuture.supplyAsync( () -> slowDown("fetching id: " + id , db.get(id)), POOL); } static ForkJoinPool POOL = new ForkJoinPool(2); public static void main(String[] args) { PipelineOfTasksExample example = new PipelineOfTasksExample(); CompletableFuture<List<String>> result = example.returnUserIdsFromDb() .thenComposeAsync(listOfIds -> { List<CompletableFuture<String>> jobs = listOfIds.parallelStream() .map(id -> example.fetchById(id)) .collect(Collectors.toList()); return CompletableFuture.allOf(jobs.toArray(new CompletableFuture<?>[0])) .thenApply(_void -> jobs.stream() .map(CompletableFuture::join).collect(Collectors.toList())); }, POOL ); System.out.println(result.join()); System.out.println(ForkJoinPool.commonPool().getPoolSize()); }}不同之处在于，首先提交所有异步作业，然后join安排调用它们的依赖操作，仅在所有作业完成后执行，因此这些join调用永远不会阻塞。只有方法join末尾的最终调用才main可能阻塞主线程。所以这会打印出类似的东西trigger building the list of Ids - thread: mainbuilding the list of Ids -> [1, 2, 3, 4] thread: ForkJoinPool-1-worker-1, 1 threadstrigger fetching id: 3 thread: ForkJoinPool-1-worker-1trigger fetching id: 2 thread: ForkJoinPool-1-worker-0trigger fetching id: 4 thread: ForkJoinPool-1-worker-1trigger fetching id: 1 thread: ForkJoinPool-1-worker-0fetching id: 4 -> user4 thread: ForkJoinPool-1-worker-1, 2 threadsfetching id: 3 -> user3 thread: ForkJoinPool-1-worker-0, 2 threadsfetching id: 2 -> user2 thread: ForkJoinPool-1-worker-1, 2 threadsfetching id: 1 -> user1 thread: ForkJoinPool-1-worker-0, 2 threads[user1, user2, user3, user4]显示无需创建补偿线程，因此线程数与配置的目标并行度相匹配。请注意，如果实际工作是在后台线程中而不是在fetchById方法本身中完成的，那么您现在不再需要并行流，因为没有阻塞join()调用。对于这种情况，仅使用stream()通常会带来更高的性能。
随时随地看视频慕课网APP
相关分类
Java