Python多进程机制实例详解

在以前只是接触过PYTHON的多线程机制，今天搜了一下多进程，相关文章好像不是特别多。看了几篇，小试了一把。程序如下，主要内容就是通过PRODUCER读一个本地文件，一行一行的放到队列中去。然后会有相应的WORKER从队列中取出这些行。

import multiprocessing
import os
import sys
import Queue
import time
def writeQ(q,obj):
    q.put(obj,True,None)
    print "put size: ",q.qsize()
def readQ(q):
    ret = q.get(True,1)
    print "get size: ",q.qsize()
    return ret
def producer(q):
    time.sleep(5)  #让进行休息几秒 方便ps命令看到相关内容
    pid = os.getpid()
    handle_file = '/home/dwapp/joe.wangh/test/multiprocess/datafile'
    with open(handle_file,'r') as f:   #with...as... 这个用法今天也是第一次看到的
        for line in f:
            print "producer <" ,pid , "> is doing: ",line
            writeQ(q,line.strip())
    q.close()
def worker(q):
    time.sleep(5)  #让进行休息几秒 方便ps命令看到相关内容
    pid = os.getpid()
    empty_count = 0
    while True:
        try:
            task = readQ(q)
            print "worker <" , pid , "> is doing: " ,task
            '''
            如果这里不休眠的话 一般情况下所有行都会被同一个子进程读取到 为了使实验效果更加清楚 在这里让每个进程读取完
一行内容时候休眠5s 这样就可以让其他的进程到队列中进行读取
            '''
            time.sleep(5)  
        except Queue.Empty:
            empty_count += 1
            if empty_count == 3:
                print "queue is empty, quit"
                q.close()
                sys.exit(0)
def main():
    cOncurrence= 3
    q = multiprocessing.Queue(10)
    funcs = [producer , worker]
    for i in range(concurrence-1):
        funcs.append(worker)
    for item in funcs:
        print str(item)
    nfuncs = range( len(funcs) )
    processes = []    
    for i in nfuncs:
        p = multiprocessing.Process(target=funcs[i] , args=(q,))
        processes.append(p)
    print "concurrence worker is : ",concurrence," working start"
    for i in nfuncs:
        processes[i].start()
    for i in nfuncs:
        processes[i].join()
    print "all DONE"
if __name__ == '__main__':
    main()

实验结果如下：

dwapp@pttest1:/home/dwapp/joe.wangh/test/multiprocess>python 1.py 




concurrence worker is : 3 working start
producer <28320 > is doing: line 1
put size: 1
producer <28320 > is doing: line 2
put size: 2
producer <28320 > is doing: line 3
put size: 3
producer <28320 > is doing: line 4
put size: 3
producer <28320 > is doing: line 5
get size: 3
put size: 4
worker <28321 > is doing: line 1
get size: 3
worker <28322 > is doing: line 2
get size: 2
worker <28323 > is doing: line 3
get size: 1
worker <28321 > is doing: line 4
get size: 0
worker <28322 > is doing: line 5
queue is empty, quit
queue is empty, quit
queue is empty, quit
all DONE

程序运行期间在另外一个窗口进行ps命令 可以观测到一些进程的信息

dwapp@pttest1:/home/dwapp/joe.wangh/test/multiprocess>ps -ef | grep python
dwapp  13735 11830 0 Nov20 pts/12  00:00:05 python
dwapp  28319 27481 8 14:04 pts/0  00:00:00 python 1.py
dwapp  28320 28319 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28321 28319 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28322 28319 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28323 28319 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28325 27849 0 14:04 pts/13  00:00:00 grep python
dwapp@pttest1:/home/dwapp/joe.wangh/test/multiprocess>ps -ef | grep python
dwapp  13735 11830 0 Nov20 pts/12  00:00:05 python     #此时28320进程 也就是PRODUCER进程已经结束
dwapp  28319 27481 1 14:04 pts/0  00:00:00 python 1.py
dwapp  28321 28319 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28322 28319 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28323 28319 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28328 27849 0 14:04 pts/13  00:00:00 grep python
dwapp@pttest1:/home/dwapp/joe.wangh/test/multiprocess>ps -ef | grep python
dwapp  13735 11830 0 Nov20 pts/12  00:00:05 python
dwapp  28319 27481 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28321 28319 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28322 28319 0 14:04 pts/0  00:00:00 python 1.py
dwapp  28323 28319 0 14:04 pts/0  00:00:00 [python]   #这里应该是代表28323进程（WORKER）已经运行结束了
dwapp  28331 27849 0 14:04 pts/13  00:00:00 grep python
dwapp@pttest1:/home/dwapp/joe.wangh/test/multiprocess>ps -ef | grep python
dwapp  13735 11830 0 Nov20 pts/12  00:00:05 python
dwapp  28337 27849 0 14:05 pts/13  00:00:00 grep python

希望本文所述对大家的Python程序设计有所帮助。