如何在 Python 中的父进程和分叉子进程之间共享数据？
声明：本页面是StackOverFlow热门问题的中英对照翻译，遵循CC BY-SA 4.0协议，如果您需要使用它，必须同样遵循CC BY-SA许可，注明原文地址和作者信息，同时你必须将它归于原作者（不是我）：StackOverFlow
原文地址: http://stackoverflow.com/questions/218935/
Warning: these are provided under cc-by-sa 4.0 license. You are free to use/share it, But you must attribute it to the original authors (not me):
StackOverFlow
How do you share data between a parent and forked child process in Python?
提问by Florian B?sch
I'm pretty sure one would do this using the os.plock(op) function, but I have no idea how. Also, if there's a better way, I'd be grateful to find out. Code snippets are very welcome.
我很确定有人会使用 os.plock(op) 函数来做到这一点，但我不知道怎么做。另外，如果有更好的方法，我将不胜感激。非常欢迎代码片段。
回答by Florian B?sch
Subprocessreplaces os.popen, os.system, os.spawn, popen2 and commands. A simple example for pipingwould be:
子进程替换了 os.popen、os.system、os.spawn、popen2 和命令。管道的一个简单示例是：
p1 = Popen(["dmesg"], stdout=PIPE)
p2 = Popen(["grep", "hda"], stdin=p1.stdout, stdout=PIPE)
output = p2.communicate()[0]
You could also use a memory mapped filewith the flag=MAP_SHARED for shared memory between processes.
您还可以使用带有 flag=MAP_SHARED的内存映射文件作为进程之间的共享内存。
multiprocessingabstracts both pipesand shared memoryand provides a higher level interface. Taken from the Processing documentation:
multiprocessing抽象了管道和共享内存，并提供了更高级别的接口。取自处理文档：
from multiprocessing import Process, Pipe
def f(conn):
conn.send([42, None, 'hello'])
conn.close()
if __name__ == '__main__':
parent_conn, child_conn = Pipe()
p = Process(target=f, args=(child_conn,))
p.start()
print parent_conn.recv() # prints "[42, None, 'hello']"
p.join()
回答by Brian
Take a look at the multiprocessingmodule new in python 2.6 (also available for earlier versions a pyprocessing
看一下python 2.6 中新的multiprocessing模块（也可用于早期版本的pyprocessing
Here's an example from the docs illustrating passing information using a pipe for instance:
这是文档中的一个示例，说明使用管道传递信息，例如：
from multiprocessing import Process, Pipe
def f(conn):
conn.send([42, None, 'hello'])
conn.close()
if __name__ == '__main__':
parent_conn, child_conn = Pipe()
p = Process(target=f, args=(child_conn,))
p.start()
print parent_conn.recv() # prints "[42, None, 'hello']"
p.join()
回答by Dan Lenski
This is pretty much Python-independent! It's a classic example of Unix interprocess communication. One good option is to use popen()to open a pipe between the parent and child processes, and pass data/messages back and forth along the pipe.
这几乎与 Python 无关！它是 Unix 进程间通信的经典示例。一个不错的选择是使用popen()在父进程和子进程之间打开管道，并沿管道来回传递数据/消息。
Take a look at the subprocessmodule, which can set up the necessary pipes automatically while spawning child processes.
看看subprocess模块，它可以在生成子进程时自动设置必要的管道。
回答by mipadi
You have two options: os.popen*in the osmodule, or you can use the subprocessmodule to the same effect. The Python manual has pretty documentation and examples for popenand subprocess.
你有两个选择：os.popen*在os模块中，或者你可以使用subprocess模块达到同样的效果。Python 手册中有关于popen和subprocess 的漂亮文档和示例。
回答by David Fraser
If you are doing low-level operating system forking and really want to avoid using pipes, it is possible to use shared memory-mapped files as well. This is not nearly as nice as using subprocessor popenpipes, but including the answer for completeness...
如果您正在执行低级操作系统分叉并且真的想避免使用管道，那么也可以使用共享内存映射文件。这几乎不像使用subprocess或popen管道那么好，但包括完整性的答案......
There's a full example here, but basically you can combine the osfile handling and mmapmodules:
这里有一个完整的例子，但基本上你可以结合os文件处理和mmap模块：
import mmap, os, tempfile
fd, tmpfile = tempfile.mkstemp()
os.write(fd, '\x00' * mmap.PAGESIZE)
os.lseek(fd, 0, os.SEEK_SET)
child_pid = os.fork()
if child_pid:
buf = mmap.mmap(fd, mmap.PAGESIZE, mmap.MAP_SHARED, mmap.PROT_READ)
os.waitpid(child_pid, 0)
child_message = buf.readline()
print(child_message)
os.close(fd)
else:
buf = mmap.mmap(fd, mmap.PAGESIZE, mmap.MAP_SHARED, mmap.PROT_WRITE)
buf.write('testing\n')
os.close(fd)
