概述

考虑这样一个问题，有hello.py脚本，输出”hello, world!”；有TestInput.py脚本，等待用户输入，然后打印用户输入的数据。那么，怎么样把hello.py输出内容发送给TestInput.py，最后TestInput.py打印接收到的”hello, world!”。下面我来逐步讲解一下shell的交互方式。

hello.py代码如下:

#!/usr/bin/pythonprint "hello, world!"

TestInput.py代码如下:

#!/usr/bin/pythonstr = raw_input()print("input string is: %s" % str)

1.os.system(cmd)

这种方式只是执行shell命令,返回一个返回码(0表示执行成功,否则表示失败)

retcode = os.system("python hello.py")print("retcode is: %s" % retcode);

输出:

hello, world!retcode is: 0

2.os.popen(cmd)

执行命令并返回该执行命令程序的输入流或输出流.该命令只能操作单向流,与shell命令单向交互,不能双向交互.
返回程序输出流,用fouput变量连接到输出流

fouput = os.popen("python hello.py")result = fouput.readlines()print("result is: %s" % result);

输出:

result is: ['hello, world!\n']

返回输入流,用finput变量连接到输出流

finput = os.popen("python TestInput.py", "w")finput.write("how are you\n")

输出:

input string is: how are you

3.利用subprocess模块

subprocess.call()

类似os.system(),注意这里的”shell=True”表示用shell执行命令,而不是用默认的os.execvp()执行.

f = call("python hello.py", shell=True)print f

输出:

hello, world!0

subprocess.Popen()

利用Popen可以是实现双向流的通信,可以将一个程序的输出流发送到另外一个程序的输入流.
Popen()是Popen类的构造函数,communicate()返回元组(stdoutdata, stderrdata).

p1 = Popen("python hello.py", stdin = None, stdout = PIPE, shell=True)p2 = Popen("python TestInput.py", stdin = p1.stdout, stdout = PIPE, shell=True)print p2.communicate()[0]#other way#print p2.stdout.readlines()

输出:

input string is: hello, world!

整合代码如下:

#!/usr/bin/pythonimport osfrom subprocess import Popen, PIPE, callretcode = os.system("python hello.py")print("retcode is: %s" % retcode);fouput = os.popen("python hello.py")result = fouput.readlines()print("result is: %s" % result);finput = os.popen("python TestInput.py", "w")finput.write("how are you\n")f = call("python hello.py", shell=True)print fp1 = Popen("python hello.py", stdin = None, stdout = PIPE, shell=True)p2 = Popen("python TestInput.py", stdin = p1.stdout, stdout = PIPE, shell=True)print p2.communicate()[0]#other way#print p2.stdout.readlines()