python: sharing variables between modules

Question

I am trying to share variables between two separate modules in python. In general what I am trying to do is,

class A:
 def __init__(self):
 self.set1 = 0
 self.set2 = 0
 def count(self, data):
 parse throught the data in here...
 self.set1 = data[0:2]
 self.set2 = data[5:20]
-------------------------------------------- 
from moduleA import A
def main():
 data = A()
 print(data.set1)
 print(data.set2)
if __name__ == '__main__': main()

but this will not work. it will just print 0 for set1 and set2. i'm not sure how I would get the correct results.

Does anyone know the right approach to solve my problem?

Thanks in advance!

• jonrsharpe: What else would it do?! You never call data.count and don't pass any data (you may have confused yourself by naming the class instance the same as a method parameter). It's hard to advise on approach when you don't even say what you're trying to do.

Answer 1

You are never calling count(...), the values of set1 and set2 will be 0 until you do.

You call __init__(self) by instantiating the class with the line:

data = A()

This sets both data.set1 and data.set2 to 0.

---

Update: following OP's comment.

When you create a new A object in your main module, you are creating a new instance of that object. Each instance has it's own version of its attributes, so if I did:

data1 = A()
data2 = A()
data1.set1 = "some value"

Then the following would be false:

data1.set1 == data2.set1

Both objects get their own set of instance attributes.

To acheive what you want, you can do two (sensible) things:

• Store the data on the class itself
• Have a singleton object that is declared in ModuleA and that all the other modules use

If will elaborate or both below.

Class Attributes

If you declare an attribute in the class definition, instead of the `init(self, ...) method, those attributes are shared between every instance, at least initially.

An example:

class A(object):
 a = []

Now if I do this:

 data1 = A()
 data2 = A()
 data1.a.append('some value')

Then data1.a and data2.a will both be the same list, with the same values.

Like I said, this is only initially, if you ever assign to a, then they will be referencing different things, and won't be the same. I could re-bind data1.a to something else, and data2.a would have the original value:

 data1.a = ["another value"]
 print(data1.a) # prints ["another value"]
 print(data2.a) # prints ["some value"]

If you want all A objects to share that attribute value, you must never assign to it for a particular instance, because then that instance will have its own value.

"Singleton" object

Instead of creating a new object data in main, you could declare a single object of type a in ModuleA:

 class A(object):
 ...
 data = A()

Now, in your main module you can use it like this:

 from ModuleA import data
 print(data.set1)

But you must always ensure that all your modules use that single object, and don't create their own As.

• user1982824: I understand that. However, if I do call count I would have to pass something in for data, but I just want to be able to access the variables inside it. the count function is being used by other modules as well, but for main() I just want to be able to access them
• Jamie Cockburn: You are then looking for a singleton(ish) pattern. I'll update my answer.
• user1982824: Even if I declare data = A() within moduleA, i still will not have access to the variables within count(self, data)
• user1982824: set1 and set2 would still have values of 0
• Jamie Cockburn: You said that "other modules" use the count function. You are not being very clear what exactly you expect to happen...