Python的多線程編程

提到多線程，很多人就會望而卻步，本文將由淺入深地帶你攻克python多線程編程，並防止你跳入深坑，

首先看一段簡單的代碼：

from time import ctime,sleep
def play_video(video):
    for i in range(2):
        print "i am playing video: %s at %s"%(video,ctime())
        sleep(4)


def play_music(music):
    for i in range(2):
        print "i am playing music: %s at %s"%(music,ctime())
        sleep(2)


if __name__=="__main__":
    
    play_video("speed_and_crazy")
    play_music("chengdu")

    print "all are over at %s"%ctime()
                
    

執行結果：

C:\Python27>python mui_thread.py
i am playing video: speed_and_crazy at Mon Jun 26 23:01:59 2017
i am playing video: speed_and_crazy at Mon Jun 26 23:02:03 2017
i am playing music: chengdu at Mon Jun 26 23:02:07 2017
i am playing music: chengdu at Mon Jun 26 23:02:09 2017
all are over at Mon Jun 26 23:02:11 2017

隨着人們對多任務的要求，同時為了充分利用cpu資源，多線程編程不可避免，那么我們如何利用python去實現play_video和play_music

兩個任務同時運行呢？

from time import ctime,sleep
import threading
def play_video(video):
    for i in range(2):
        print "i am playing video: %s at %s \n"%(video,ctime())
        sleep(5)


def play_music(music):
    for i in range(2):
        print "i am playing music: %s at %s \n"%(music,ctime())
        sleep(1)

threads=[]

thread1=threading.Thread(target=play_video,args=("speed_and_crazy",))

threads.append(thread1)

thread2=threading.Thread(target=play_music,args=("chengdu",))

threads.append(thread2)



if __name__=="__main__":

    for thread in threads:
        thread.setDaemon(True) #將線程聲明為守護線程，必須在start()方法調用之前，如果不設置為守護線程,程序會被無限掛起
        thread.start()

    print "all are over at %s \n"%ctime()
                
    

測試結果：

C:\Python27>python mui_thread.py
i am playing video: speed_and_crazy at Mon Jun 26 23:18:52 2017
all are over at Mon Jun 26 23:18:52 2017
i am playing music: chengdu at Mon Jun 26 23:18:52 2017  #從打印的時間可知，play_video、play_music和父進程幾乎同時運行

從結果看，與我們最初的目標相差甚遠，怎么沒有按照順序執行，為什么每個函數都只有一條日記輸出？

那是因為子線程（play_video、play_music）和主線程print "all are over at %s \n"%ctime()都是同一時間啟動，但由於主線程已經運行結束，所以導致子線程也同時終止，在這種條件下，我們如何保證子進程都能夠執行完畢呢？

增加thread.join()並 放在循環外

from time import ctime,sleep
import threading
def play_video(video):
    for i in range(2):
        print "i am playing video: %s at %s \n"%(video,ctime())
        sleep(1)


def play_music(music):
    for i in range(2):
        print "i am playing music: %s at %s \n"%(music,ctime())
        sleep(5)

threads=[]

thread1=threading.Thread(target=play_video,args=("speed_and_crazy",))

threads.append(thread1)

thread2=threading.Thread(target=play_music,args=("chengdu",))

threads.append(thread2)



if __name__=="__main__":

    for thread in threads:
        thread.setDaemon(True)
        thread.start()
    thread.join()  #加在循環外，
    print "all are over at %s \n"%ctime()
                
    

運行結果：

C:\Python27>python mui_thread.py
i am playing video: speed_and_crazy at Mon Jun 26 23:32:21 2017
i am playing music: chengdu at Mon Jun 26 23:32:21 2017


i am playing video: speed_and_crazy at Mon Jun 26 23:32:22 2017

i am playing music: chengdu at Mon Jun 26 23:32:26 2017

all are over at Mon Jun 26 23:32:31 2017

thread.join()的作用是主線程必須等待子線程都執行完了才能結束，play_video、play_music幾乎同時執行
但是如果改變play_video、play_music里面的sleep的時長，即是下面的代碼：

from time import ctime,sleep
import threading
def play_video(video):
    for i in range(2):
        print "i am playing video: %s at %s \n"%(video,ctime())
        sleep(5)


def play_music(music):
    for i in range(2):
        print "i am playing music: %s at %s \n"%(music,ctime())
        sleep(1)

threads=[]

thread1=threading.Thread(target=play_video,args=("speed_and_crazy",))

threads.append(thread1)

thread2=threading.Thread(target=play_music,args=("chengdu",))

threads.append(thread2)



if __name__=="__main__":

    for thread in threads:
        thread.setDaemon(True)
        thread.start()
    thread.join()
    print "all are over at %s \n"%ctime()
                
    

此時運行結果：

C:\Python27>python mui_thread.py
i am playing video: speed_and_crazy at Mon Jun 26 23:44:13 2017
i am playing music: chengdu at Mon Jun 26 23:44:13 2017


i am playing music: chengdu at Mon Jun 26 23:44:14 2017

all are over at Mon Jun 26 23:44:15 2017

我們看到play_video還有一條log沒有打印出來，原因是thread.join()在循環外，此時的thread為play_music，父進程只會等待play_music進程執行完就結束，而不會等待play_video（sleep時間較長）執行完才結束，所以才會有上面的結果