Python线程中的全局变量与参数

Question

I am trying to update a dictionary and read a list using multiple threads so which one is better, is it better to pass the list and dictionary as parameters in the thread or use them in the thread using the global keyword.

ie.

mythread(list, list_semaphore, dict, dict_semaphore)

vs

mythread():
    global dict
    global dict_semaphore
    global list
    global list_semaphore

Answer 1

Is not just a matter of threading , but by using parameters your function will work for any combination of parameters data, not just some specific global variables.

Lets say you have this:

def foo(lst, lst_semaphore, dct, dct_semaphore):
    do_some_nice_stuff()

You can spam any thread doind that with any list or dict:

threading.Thread(target=foo, args = (lst1, lst_semaphore1, dct1, dct_semaphore1))
threading.Thread(target=foo, args = (lst2, lst_semaphore2, dct1, dct_semaphore1))
threading.Thread(target=foo, args = (lst2, lst_semaphore2, dct2, dct_semaphore2))

If you use globals you will have to redefine a function for each combination:

def foo():
    global lst1
    global lst_semaphore1
    global dct1
    global dct_semaphore1
    do_nice_stuff()
...
def foo2():
    global lst2
    global lst_semaphore2
    global dct2
    global dct_semaphore2
    do_nice_stuff()


threading.Thread(target=foo)
threading.Thread(target=foo1)
threading.Thread(target=foo2)

So, using parameters will make your code more reusable mostly always.