首页 > 代码库 > Python 中模块间全局变量的使用上的注意

Python 中模块间全局变量的使用上的注意

最近用Python写代码,需要用到模块间的全局变量。

网上四处搜索,发现普遍做法是把全局变量放到一个独立的模块中,使用时,导入此全局变量模块即可。

但是在实际使用过程中发现了些小问题:在使用如下代码导入全局变量模块时,各个模块获取到的全局变量都是初始值。

from module import global_var

但是如果使用「模块名.全局变量名」来访问时,却又是正常的:

import moduleprint module.global_var

为了弄清其中的原委,我写了个测试程序来仔细查看其中的细节:

 

 

1. import module

global_var.py

GLOBAL_VAR = [1, 2, 3]

m1.py

import global_varimport m2print m1: before appending:          , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VARglobal_var.GLOBAL_VAR.append(m1)print m1: after appending:           , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VARprint m1: before calling m2.append():, id(global_var.GLOBAL_VAR), global_var.GLOBAL_VARm2.append()print m1: after calling m2.append(): , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VARprint -----------------print m1: before assigning:          , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VARglobal_var.GLOBAL_VAR = [m1]print m1: after assigning:           , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VARprint m1: before calling m2.assign():, id(global_var.GLOBAL_VAR), global_var.GLOBAL_VARm2.assign()print m1: after calling m2.assign(): , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VAR

m2.py

import global_vardef append():    print m2: before assiging:           , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VAR    global_var.GLOBAL_VAR.append(m2)    print m2: after assiging:            , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VARdef assign():    print m2: before assiging:           , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VAR    global_var.GLOBAL_VAR = [m2]    print m2: after assiging:            , id(global_var.GLOBAL_VAR), global_var.GLOBAL_VAR

运行结果:

Mac-pastgift:pytest pastgift$ python global_test_import/m1.pym1: before appending:           4457308312 [1, 2, 3]m1: after appending:            4457308312 [1, 2, 3, ‘m1‘]m1: before calling m2.append(): 4457308312 [1, 2, 3, ‘m1‘]m2: before assiging:            4457308312 [1, 2, 3, ‘m1‘]m2: after assiging:             4457308312 [1, 2, 3, ‘m1‘, ‘m2‘]m1: after calling m2.append():  4457308312 [1, 2, 3, ‘m1‘, ‘m2‘]-----------------m1: before assigning:           4457308312 [1, 2, 3, ‘m1‘, ‘m2‘]m1: after assigning:            4457444848 [‘m1‘]m1: before calling m2.assign(): 4457444848 [‘m1‘]m2: before assiging:            4457444848 [‘m1‘]m2: after assiging:             4457308312 [‘m2‘]m1: after calling m2.assign():  4457308312 [‘m2‘]

在这种方式下,如果是改变对象本身的操作(append),各个模块内部的操作都是针对于同一个对象。

而赋值操作,虽然全局变量所指向的对象改变了,但是全局变量本身依然在各个模块内部能够正确访问,这正是我所希望的效果。

 

 

2. from module import GLOBAL_VAR

global_var.py

GLOBAL_VAR = [1, 2, 3]

m1.py

from global_var import GLOBAL_VARimport m2print m1: before appending:          , id(GLOBAL_VAR), GLOBAL_VARGLOBAL_VAR.append(m1)print m1: after appending:           , id(GLOBAL_VAR), GLOBAL_VARprint m1: before calling m2.append():, id(GLOBAL_VAR), GLOBAL_VARm2.append()print m1: after calling m2.append(): , id(GLOBAL_VAR), GLOBAL_VARprint -----------------print m1: before assigning:          , id(GLOBAL_VAR), GLOBAL_VARGLOBAL_VAR = [m1]print m1: after assigning:           , id(GLOBAL_VAR), GLOBAL_VARprint m1: before calling m2.assign():, id(GLOBAL_VAR), GLOBAL_VARm2.assign()print m1: after calling m2.assign(): , id(GLOBAL_VAR), GLOBAL_VAR

m2.py

from global_var import GLOBAL_VARdef append():    global GLOBAL_VAR    print m2: before assiging:           , id(GLOBAL_VAR), GLOBAL_VAR    GLOBAL_VAR.append(m2)    print m2: after assiging:            , id(GLOBAL_VAR), GLOBAL_VARdef assign():    global GLOBAL_VAR    print m2: before assiging:           , id(GLOBAL_VAR), GLOBAL_VAR    GLOBAL_VAR = [m2]    print m2: after assiging:            , id(GLOBAL_VAR), GLOBAL_VAR

运行结果:

Mac-pastgift:pytest pastgift$ python global_test_from_import/m1.pym1: before appending:           4539998360 [1, 2, 3]m1: after appending:            4539998360 [1, 2, 3, ‘m1‘]m1: before calling m2.append(): 4539998360 [1, 2, 3, ‘m1‘]m2: before assiging:            4539998360 [1, 2, 3, ‘m1‘]m2: after assiging:             4539998360 [1, 2, 3, ‘m1‘, ‘m2‘]m1: after calling m2.append():  4539998360 [1, 2, 3, ‘m1‘, ‘m2‘]-----------------m1: before assigning:           4539998360 [1, 2, 3, ‘m1‘, ‘m2‘]m1: after assigning:            4540135112 [‘m1‘]m1: before calling m2.assign(): 4540135112 [‘m1‘]m2: before assiging:            4539998360 [1, 2, 3, ‘m1‘, ‘m2‘]m2: after assiging:             4540135040 [‘m2‘]m1: after calling m2.assign():  4540135112 [‘m1‘]

这次,运行结果和上次略有不同。

改变对象本身的操作(append)和之前的例子一样,全局变量始终指向相同的对象。

但是赋值操作就比较奇怪了。注意标红部分,每个模块第一次获取到的这个「全局变量」都是相同id的对象。即使这个「全局变量」在其他模块中已经被重新赋值,但在本模块中依然指向的是最原始的id。

显然,这个算不上是「全局变量」了。

 

Python 中模块间全局变量的使用上的注意