这个想法是运行一组任务，当它们完成时处理它们，一旦返回所需的结果，就取消剩余的任务。我不介意等待已经开始的任务完成，但应该禁止开始新的任务。
我试过调用pool.shutdown(cancel_futures=True)，但它似乎不工作在这里：

import time
import multiprocessing
import concurrent.futures
import random
filelist = ['001', '002', '003', '004', '005', '006', '007', '008', '009', '010']
parallel = 2
def worker(name, lock, index):
    with lock:
        index.value += 1
        pc = index.value*10
        print(f'starting: {name} ({pc})')
    d = random.randrange(1, 5)
    time.sleep(d)
    return pc == 60
print('INIT')
with multiprocessing.Manager() as manager:
    lock = manager.Lock()
    index = manager.Value('b', 0)
    if parallel > 0:
        with concurrent.futures.ProcessPoolExecutor(max_workers=parallel) as pool:
            tasks = [pool.submit(worker, filename, lock, index) for filename in filelist]
            for task in concurrent.futures.as_completed(tasks):
                rc = task.result()
                if rc:
                    print('Found!')
                    pool.shutdown(cancel_futures=True)
                    break
      else:
          for filename in filelist:
              rc = worker(filename, lock, index)
              if rc:
                  print('Found!')
                  break
print('END')

字符串
在串行模式（parallel = 0）下运行时，我得到了所需的结果：

INIT
starting: 001 (10)
starting: 002 (20)
starting: 003 (30)
starting: 004 (40)
starting: 005 (50)
starting: 006 (60)
Found!
END

型
但当并行启动时，它会一直持续到最后：

INIT
starting: 001 (10)
starting: 002 (20)
starting: 003 (30)
starting: 004 (40)
starting: 005 (50)
starting: 006 (60)
Found!
starting: 007 (70)
starting: 008 (80)
starting: 009 (90)
END

型
那么，有没有办法退出as_completed循环呢？