手把手带你理解python3多线程【1】

写在最前面：python3使用两种模块引入线程的概念

• _thread
• threading(推荐使用)

由于在python2中使用thread模块，而在python3中该模块已经被废弃，使用_thread来兼容，今天的例子都是基于python3的threading

 

首先，我们来看一下没有用线程的例子：

import time
import threading
from time import ctime,sleep


def First():
    for i in range(3):
        print("The First start at %s" % ctime())
        sleep(1)


def Second():
    for i in range(3):
        print("The Second start at %s" % ctime())
        sleep(5)


if __name__ == '__main__':
    First()
    Second()
    print("all over %s" %ctime())

The First start at Mon Oct 29 16:41:26 2018
The First start at Mon Oct 29 16:41:27 2018
The First start at Mon Oct 29 16:41:28 2018
The Second start at Mon Oct 29 16:41:29 2018
The Second start at Mon Oct 29 16:41:34 2018
The Second start at Mon Oct 29 16:41:39 2018
Exception ignored in: <module 'threading' from 'F:\\leetcode\\threading.py'>
AttributeError: module 'threading' has no attribute '_shutdown'
all over Mon Oct 29 16:41:44 2018

Process finished with exit code 0

一个小插曲，出现了一个报错，这是因为我把这个py文件命名为threading.py和线程本身的模块名重复了

上面的例子很好理解，程序很听话的按照我们调用函数的顺序执行了First，Second，那么有没有办法同时执行呢？

可以，通过threading线程

为了更好的理解threading，我们在启用线程的时候加入参数

import time
import threading
from time import ctime,sleep


def First(name):
    for i in range(3):
        print("The %s start at %s" % (name, ctime()))
        sleep(1)


def Second(name):
    for i in range(3):
        print("The %s start at %s" % (name, ctime()))
        sleep(5)

threads = []
t1 = threading.Thread(target=First,args=(u'First',))
threads.append(t1)
t2 = threading.Thread(target=Second,args=(u'Second',))
threads.append(t2)

if __name__ == '__main__':
    for t in threads:
        t.setDaemon(True)
        t.start()
    print("all over %s" %ctime())

这时的输出结果是：

The First start at Mon Oct 29 17:03:44 2018
The Second start at Mon Oct 29 17:03:44 2018
all over Mon Oct 29 17:03:44 2018

Process finished with exit code 0

大家有没有发现，我们明明在子线程中写了循环遍历，为什么输出只有一个？

这里我们讲一讲setDaemon

setDaemom默认false，当我们置为True时，子线程为守护线程，主线程一旦执行结束，则全部子线程被强制终止。所以当for循环结束，所有的子线程被起来后，主线程也结束，此时子线程强制结束。

那如何使得子线程继续执行呢？

import time
import threading
from time import ctime,sleep


def First(name):
    for i in range(3):
        print("The %s start at %s" % (name, ctime()))
        sleep(1)


def Second(name):
    for i in range(3):
        print("The %s start at %s" % (name, ctime()))
        sleep(5)

threads = []
t1 = threading.Thread(target=First,args=(u'First',))
threads.append(t1)
t2 = threading.Thread(target=Second,args=(u'Second',))
threads.append(t2)

if __name__ == '__main__':
    for t in threads:
        #t.setDaemon(True)
        t.start()
    print("all over %s" %ctime())

此时的结果如下：

The First start at Mon Oct 29 17:08:57 2018
The Second start at Mon Oct 29 17:08:57 2018
all over Mon Oct 29 17:08:57 2018
The First start at Mon Oct 29 17:08:58 2018
The First start at Mon Oct 29 17:08:59 2018
The Second start at Mon Oct 29 17:09:02 2018
The Second start at Mon Oct 29 17:09:07 2018

Process finished with exit code 0

大家可以看到，主线程结束后，子线程继续执行

 

那么有没有办法使得主线程等待子线程执行完毕再执行呢？

可以，使用join()

import time
import threading
from time import ctime,sleep


def First(name):
    for i in range(3):
        print("The %s start at %s" % (name, ctime()))
        sleep(1)


def Second(name):
    for i in range(3):
        print("The %s start at %s" % (name, ctime()))
        sleep(5)

threads = []
t1 = threading.Thread(target=First,args=(u'First',))
threads.append(t1)
t2 = threading.Thread(target=Second,args=(u'Second',))
threads.append(t2)

if __name__ == '__main__':
    for t in threads:
        t.setDaemon(True)
        t.start()
    t.join()
    print("all over %s" %ctime())

这时我们可以清晰的看到

The First start at Mon Oct 29 17:11:14 2018
The Second start at Mon Oct 29 17:11:14 2018
The First start at Mon Oct 29 17:11:15 2018
The First start at Mon Oct 29 17:11:16 2018
The Second start at Mon Oct 29 17:11:19 2018
The Second start at Mon Oct 29 17:11:24 2018
all over Mon Oct 29 17:11:29 2018

Process finished with exit code 0

当所有子线程都结束之后，主线程才开始执行，此时的子线程不需要守护，因为我们使用join()保证了只有当子线程全部结束，主线程才会执行

那如果我们把join()加入for循环里面呢？

The First start at Mon Oct 29 17:29:36 2018
The First start at Mon Oct 29 17:29:37 2018
The First start at Mon Oct 29 17:29:38 2018
The Second start at Mon Oct 29 17:29:39 2018
The Second start at Mon Oct 29 17:29:44 2018
The Second start at Mon Oct 29 17:29:49 2018
all over Mon Oct 29 17:29:54 2018

Process finished with exit code 0

此时，当第一个子线程还没结束时，主线程无法再次调用for循环起第二个子线程，只能等待第一个子线程结束后，开始第二个子线程。

 

今天头疼，先讲到这儿，最优化和C++算法作业还没写，难受