PythonI/O进阶学习笔记_10.python的多线程(4)

当前位置:

首页 > Python基础教程 >

PythonI/O进阶学习笔记_10.python的多线程(4)

3 successed! 3 2

为什么done()输出的是false呢。因为submit的返回是非阻塞的，没有等task执行完就返回了task done的状态。

如果隔几秒输出done()的返回又是true了。

Future对象常用的其他方法：cancel()

取消该 Future 代表的线程任务。如果该任务正在执行，不可取消，则该方法返回 False；否则，程序会取消该任务，并返回 True。

print(task1.done())
print(task1.cancel())
print(task2.done())

#result 是阻塞的，接受函数的返回
print(task1.result())

print(task2.result())

output:
False
False
False
get html page 2 successed!
get html page 3 successed!
3
2

cancelled()：返回 Future 代表的线程任务是否被成功取消。

2.获取所有完成的future的状态和值as_completed和map

2.1 .as_complete

from concurrent.futures import  ThreadPoolExecutor,as_completed
import time
 
def get_html(uid):
    time.sleep(uid)
    url="www.test.com/{}".format(uid)
    print("get url successed: \" {} \"".format(url))
    return uid
excutor=ThreadPoolExecutor(max_workers=2)
uids=[5,2,3]
future_list=[ excutor.submit(get_html,(uid)) for uid in uids]
for future in as_completed(future_list):
    print(future.result())
 
#output:
get url successed: " www.test.com/2 "
2
get url successed: " www.test.com/5 "
5
get url successed: " www.test.com/3 "
3

as_completed():yield 所有完成的futures的所有返回。

那么as_complete是如何做到收集所有完成的异步方法的状态的呢？

先把所有已经是finish状态的future返回，

再一直while pending，等待timeout范围内的future变成finish，把finish的future yield出来。

from concurrent.futures import  ThreadPoolExecutor,as_completed
import time
 
def get_html(uid):
    time.sleep(uid)
    url="www.test.com/{}".format(uid)
    print("get url successed: \" {} \"".format(url))
    return uid
excutor=ThreadPoolExecutor(max_workers=2)
uids=[5,2,3]
future_list=[ excutor.submit(get_html,(uid)) for uid in uids]
for future in as_completed(future_list):
    print(future.result())
 
#output:
get url successed: " www.test.com/2 "
2
get url successed: " www.test.com/5 "
5
get url successed: " www.test.com/3 "
3

2.2 通过excutor的map方法获取已经完成的future

excutor的map，和map是差不多的，传递函数和参数列表，就会多多线程运行参数列表数的线程。

与a中不一样的是，map返回的顺序和url中的顺序是一样的，而a的as_completed是谁先finishi谁就先被yield出来。

而且map返回的就是result，而as_completed返回的是Future。

from concurrent.futures import  ThreadPoolExecutor,as_completed,wait
import time
def get_html(uid):
    time.sleep(uid)
    url="www.test.com/{}".format(uid)
    print("get url successed: \" {} \"".format(url))
    return uid
 
excutor=ThreadPoolExecutor(max_workers=2)
uids=[5,2,3]
 
result_list=excutor.map(get_html,uids)
for result in result_list:
    print(result)
 
#output:
get url successed: " www.test.com/2 "
get url successed: " www.test.com/5 "
5
2
get url successed: " www.test.com/3 "
3

3.wait()方法

wait方法用于阻塞，指定某一个task或者一些task执行完成再往下执行。

def wait(fs, timeout=None, return_when=ALL_COMPLETED)

例：如果我想在全部task执行完之后打印"task end"字符串

from concurrent.futures import  ThreadPoolExecutor,as_completed,wait
import time
def get_html(uid):
    time.sleep(uid)
    url="www.test.com/{}".format(uid)
    print("get url successed: \" {} \"".format(url))
    return uid
excutor=ThreadPoolExecutor(max_workers=2)
uids=[5,2,3]
future_list=[ excutor.submit(get_html,(uid)) for uid in uids]
print("task end")
 
#output:
task end
get url successed: " www.test.com/2 "
get url successed: " www.test.com/5 "
get url successed: " www.test.com/3 "

#还没有执行完，就输出了 task end。需要加上：
wait(future_list)
print("task end")
 
#output:
get url successed: " www.test.com/2 "
get url successed: " www.test.com/5 "
get url successed: " www.test.com/3 "
task end

wait还可以指定在什么时候执行完后返回。

栏目列表