Question

我正在编写一个我希望监控其状态的多处理过程。如何从thaht context访问my_var？

from multiprocessing import Process
import time

my_var = list()


def alter_my_var():
    global my_var
    for x in range(10):
        my_var.append(x)
        time.sleep(1)

p = Process(target=alter_my_var)
p.start()

while p.is_alive():
    print "Length of my_var is %i" % len(my_var)
    time.sleep(1)

p.join()

print "Done - final length of my_var is %s" % len(my_var)

由于

Answer 1

您使用的是multiprocessing而不是threading。进程在不同的内存空间中运行，变量被复制到子进程，但它们不再指向相同的内存地址。因此，主进程无法读取您在进程中修改的内容。

如果您不关心某个流程或某个主题，请尝试按Process更改Thread，这样就可以了。

<强>输出：

Length of my_var is 0
Length of my_var is 1
Length of my_var is 2
Length of my_var is 3
Length of my_var is 4
Length of my_var is 6
Length of my_var is 7
Length of my_var is 8
Length of my_var is 9
Length of my_var is 10
Done - final length of my_var is 10

修改

正如其他人所说，如果你想继续使用一个过程，你必须建立某种IPC。您可以使用该队列，管理器等。

Answer 2

使用Manager，每个进程都会获得列表的副本，因此您不会与 Manager（）。list（）共享一个对象：

from multiprocessing import Process, Manager
import time

my_var = Manager().list()

def alter_my_var(my_var):
    for x in range(10):
        my_var.append(x)
        time.sleep(1)

p = Process(target=alter_my_var, args=(my_var, ))
p.start()
while p.is_alive():
    print "Length of my_var is %i" % len(my_var)
    time.sleep(1)
p.join()
print "Done - final length of my_var is %s" % len(my_var)

如果我们将代码粘贴到Ipython中，您可以看到输出：

## -- End pasted text --
Length of my_var is 0
Length of my_var is 1
Length of my_var is 3
Length of my_var is 4
Length of my_var is 5
Length of my_var is 6
Length of my_var is 7
Length of my_var is 8
Length of my_var is 9
Length of my_var is 10
Done - final length of my_var is 10

Answer 3

试试这个：

from multiprocessing import Process, Pipe
import time

my_var = list()


def alter_my_var(my_var, conn):
    for x in range(10):
        my_var.append(x)
        conn.send(len(my_var))
        time.sleep(1)

global my_var
parent_conn, child_conn = Pipe()
p = Process(target=alter_my_var, args=(my_var,child_conn,))
p.start()

while p.is_alive():
    myvar_len = parent_conn.recv()
    print "Length of my_var is %i" % myvar_len       
    time.sleep(1)

p.join()

print "Done - final length of my_var is %s" % myvar_len

访问全局变量以跟踪多处理中的状态

3 个答案: