请教python multiprocessing的问题

Question

main.py

import multiprocessing.pool as Pool
from test import set_data, get_multiprocessing



if __name__ == '__main__':
    data_list = ['data1', 'data2', 'data3']
    set_data(data_list)
    res = get_multiprocessing(2)
    print(res)

test.py

import time
import multiprocessing

_t = None

def set_data(t):
    time.sleep(1)
    global _t
    _t = t

def get_data(idx):

    return _t[idx]

def get_multiprocessing(num_threads):
    pool = multiprocessing.Pool(processes=num_threads)
    res = pool.map(get_data, range(1))
    pool.close()
    pool.join()

    return [r for r in res]

我想要等这个全局变量赋值之后再进行多进程，有什么办法吗？

向大佬们请教

Answer 1

和线程不同，进程间不能共享全局变量，也不共享对象，进程间只能通过共享内存、管道、信号量、事件等手段通讯。如果题主一定要进程间共享这个全局变量，恐怕只能像下面的代码一样将全局变量作为参数传进去——这已经不是共享对象，而是副本传参了。

import time
import multiprocessing as mp

_t = None

def set_data(t):
    time.sleep(1)
    global _t
    _t = t

def get_data(arg):
    arr, idx = arg
    return arr[idx]

def get_multiprocessing(num_threads):
    pool = mp.Pool(processes=num_threads)
    res = pool.map(get_data, ((_t,0),(_t,1),(_t,2)))
    pool.close()
    pool.join()

    return list(res)

if __name__ == '__main__':
    data_list = ['data1', 'data2', 'data3']
    set_data(data_list)
    res = get_multiprocessing(2)
    print(res)