OK I am doing some threading, and I guess when I started doing threading I assumed you can't return values like a definition (its the end of the day and my brain is about to die so maybe this is incorrect and I should start going back to get rid of global variables)
Anyway I have a test program to figure out why I can't modularize my code
a file called config.py
a_variable=0
a file called test_for_sean.py
from config import *
def blah():
global a_variable
a_variable=14
return 0
a file called main.py
from config import *
from test_for_sean import *
print a_variable #this prints correctly
blah()
print a_variable #this is still printing 0....
someone link me to something so I don't kill myself
3 Answers 3
When you import variables from a module the names are imported into the current module's namespace, they are not shared between both modules. The global keyword does not allow you to share names between modules, it only allows you to assign to a name that is in the global scope for that module.
To actually share variables between modules you need to access the variable through its module:
a file called config.py
a_variable=0
a file called test_for_sean.py
import config
def blah():
config.a_variable=14
return 0
a file called main.py
import config
from test_for_sean import blah
print config.a_variable # prints 0
blah()
print config.a_variable # prints 14
Comments
Try these changes,
config.py,
a_variable = 0
tfs.py (was test_for_sean),
import config
def blah():
config.a_variable = 14
return 0
main.py,
import tfs, config
print config.a_variable
tfs.blah()
print config.a_variable
We still import everything from config, but are 'global' variables stay in their own modules. This way we can have global variables but still let main.py define its own a_variable if it needs to.
2 Comments
There is an analogy with a filesystem. It is not a perfect analogy, but helps to understand the problem quickly.
Modules are like directories holding files and these files represent variables and other objects. In modules you are using relative paths. Just write down the absolute paths and the picture will be clear.
Your starting point is:
/my_app/module1/
/my_app/module1/a_variable
/my_app/module2/
A. If you do this in module2:
import a_variable from module1
You are making a copy of the variable a_variable:
/my_app/module1/
/my_app/module1/a_variable
/my_app/module2/
/my_app/module2/a_variable
In this case, when the variables are not mutable, they will get out of sync at the first modification.
B. But when you do this instead:
import module1
You are creating a copy of the path leading to the variable: module1.a_variable
/my_app/module1/
/my_app/module1/a_variable
/my_app/module2/
/my_app/module2/module1 -> ../module1
The result is the same variable in both modules.
Comments
Explore related questions
See similar questions with these tags.
import *. Perhapsimport test_for_sean as tfs. This will cause all sorts of problems -- like how config.py is executed twice and it fills the global namespace of both scripts its imported into.printin config.py and it only gets run once.from x import *:)