Python多进程机制实例详解
(编辑:jimmy 日期: 2024/11/18 浏览:3 次 )
本文实例讲述了Python多进程机制。分享给大家供大家参考。具体如下:
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 <function producer at 0xb7b9141c> <function worker at 0xb7b91454> <function worker at 0xb7b91454> <function worker at 0xb7b91454> 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] <defunct> #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。
下一篇:python简单判断序列是否为空的方法