I am struggling to understand the following behavior of global variables from a module. I want to import a variable from another module, and change its value through a function call (this is because in my use case that function had to access some other 'private' parts of the module). I know how I can avoid to use globals, I did it, but I am still curious.
First method
This is the preferred method following the python documentation. All globals are put in an external config.py which has to be imported everywhere needed.
#config.py
s = 'initialized'
#module.py
import config
def chg():
config.s = 'changed'
prt()
def prt():
print(config.s)
#main.py
import config
import module
print(config.s)
module.chg()
print(config.s)
This works fine (I also read some threads here about this method) : not only we can access s from other files but it is possible to change its value from everywhere, and this change is "persistent".
guest@desktop /tmp $ python3 main.py
initialized
changed
changed
Second method
I wasn't clear about whether it was needed or just highly recommended to separated all global stuff into another file.
#module.py
s = 'initialized'
def chg():
global s
s = 'changed'
prt()
def prt():
print(s)
#main.py
import module
print(module.s)
module.chg()
print(module.s)
Obviously, it is not mandatory. SO this is no magic ;-)
guest@desktop /tmp $ python3 main.py
initialized
changed
changed
Third method
This is where I am lost. I take the same basis as the second one, but I change the way the global variable is imported.
#main.py
from module import s, chg
print(s)
chg()
print(s)
This does not work any more (I mean, as I expected...).
guest@desktop /tmp $ python3 main.py
initialized
changed
initialized
So, what am I missing here ? Most of the threads I read were dealing with importing a global variable but not changing its value. Both methods were shown. Thanks !
