What is the difference between initializing a variable as global var
or calling globals().update(var)
.
Thanks
When you say
global var
you are telling Python that var is the same var that was defined in a global context. You would use it in the following way:
var=0
def f():
global var
var=1
f()
print(var)
# 1 <---- the var outside the "def f" block is affected by calling f()
Without the global statement, the var inside the "def f" block would be a local variable, and setting its value would have no effect on the var outside the "def f" block.
var=0
def f():
var=1
f()
print(var)
# 0 <---- the var outside the "def f" block is unaffected
When you say globals.update(var) I am guessing you actually mean globals().update(var). Let's break it apart.
globals() returns a dict object. The dict's keys are the names of objects, and the dict's values are the associated object's values.
Every dict has a method called "update". So globals().update() is a call to this method. The update method expects at least one argument, and that argument is expected to be a dict. If you tell Python
globals().update(var)
then var had better be a dict, and you are telling Python to update the globals() dict with the contents of the var dict.
For example:
#!/usr/bin/env python
# Here is the original globals() dict
print(globals())
# {'__builtins__': <module '__builtin__' (built-in)>, '__name__': '__main__', '__file__': '/home/unutbu/pybin/test.py', '__doc__': None}
var={'x':'Howdy'}
globals().update(var)
# Now the globals() dict contains both var and 'x'
print(globals())
# {'var': {'x': 'Howdy'}, 'x': 'Howdy', '__builtins__': <module '__builtin__' (built-in)>, '__name__': '__main__', '__file__': '/home/unutbu/pybin/test.py', '__doc__': None}
# Lo and behold, you've defined x without saying x='Howdy' !
print(x)
Howdy