How to use local variable in a function and return it?

Question

I am trying to create a script that sets a local variable, references it from a function, and can return the manipulated value back to the main scope (or whatever it's called; I'm new to Python)

I have simplified my code to show the utmost basics of what I am trying to accomplish, which is to import a local from the module into a function block, I think.

I have gotten this to work by using globals, but that isn't the best solution . . .

chambersinreactor = 0;
cardsdiscarded = 0;

def find_chamber_discard(): 
    """Find chambers and discard in row (reads each player slot)"""
    chambersinreactor = 0; # Resets the variable, not what I want
    cardsdiscarded = 0; # Resets the variable, not what I want
    chambersinreactor += 1
    cardsdiscarded += 1
    return # Don't know what to put here

find_chamber_discard()

print chambersinreactor # prints as 0, should be 1
print cardsdiscarded    # prints as 0, should be 1

Solution 1

Functions shouldn't have to know what scope they're called from; the point of a function is to make a re-usable block of code that can be invoked multiple times from different places.

You communicate information to a function by passing it through its input variables. The function communicates information back to its caller by returning it.

Managing the variables of a scope is the job of the code in that scope not any functions it invokes. If you need to set variables to values determined by a function, then you have the function return those values and you use them to set the variables. If the values the function is calculating depend on the values of variables you have in the calling scope, then you need to pass them to the function as arguments. The function you're calling shouldn't have to know what variables you're using, and shouldn't be able to mess with them.

Putting that all together, what you want to do is something like this:

def find_chamber_discard(chambersinreactor, cardsdiscarded):
    chambersinreactor += 1
    cardsdiscarded += 1
    return (chambersinreactor, cardsdiscarded)

chambersinreactor = 0;
cardsdiscarded = 0;

chambersinreactor, cardsdiscarded = find_chamber_discard(chambersinreactor, cardsdiscarded)

print chambersinreactor
print cardsdiscarded

There are ways to get around this with global variables or manipulating mutable data structures, but ultimately they make your program less flexible and more likely to contain errors that will be difficult to spot. There is a place for those techniques, but the first method you reach for to communicate information to and from functions really should be passing arguments and receiving return values.

Solution 2

#!/usr/bin/env python
chambersinreactor = 0; cardsdiscarded = 0;

def find_chamber_discard():
    chambersinreactor = 0
    cardsdiscarded = 0
    chambersinreactor += 1 
    cardsdiscarded += 1 
    return(chambersinreactor, cardsdiscarded)

#Here the globals remain unchanged by the locals.
#In python, a scope is similar to a 'namespace'
find_chamber_discard()

print chambersinreactor #prints as 0
print cardsdiscarded 

#I've modified the function to return a pair (tuple) of numbers.
#Above I ignored them. Now I'm going to assign the variables in the 
#main name space to the result of the function call.
print("=====with assignment===")
(chambersinreactor, cardsdiscarded) = find_chamber_discard()

print chambersinreactor #  now prints as 1
print cardsdiscarded 

# Here is another way that doesn't depend on returning values.
#Pass a dictionary of all the main variables into your function
#and directly access them from within the function as needed

print("=======using locals===")
def find_chamber_discard2(_locals):
    _locals['chambersinreactor'] += 1
    _locals['cardsdiscarded'] += 1
    return

find_chamber_discard2(locals())

print chambersinreactor #incremented the value, which was already 1
print cardsdiscarded 

Solution 3

One approach is to use mutable values, like dicts or lists:

settings = dict(
    chambersinreactor = 0,
    cardsdiscarded = 0
)

def find_chamber_discard():
    settings['chambersinreactor'] += 1
    settings['cardsdiscarded'] += 1

find_chamber_discard()

print settings['chambersinreactor']
print settings['cardsdiscarded']

However, if you have a function that is changing some state, you're probably better off wrapping that all up in class, as that's what they're for:

class CardCounter(object):
    def __init__(self):
        chambersinreactor = 0
        cardsdiscarded = 0

    def find_chamber_discard(self, hand):
        for card in hand:
            if card.is_chamber:
                self.chambersinreactor += 1
            if card.is_discarded:
                self.cardsdiscarded += 1

If what you're doing is counting, maybe you could use Counter:

from collections import Counter

def test_for_chamberness(x): return x == 'C'
def test_for_discarded(x): return x == 'D'

def chamber_or_discard(card):
    if test_for_chamberness(card):
        return 'chambersinreactor'
    if test_for_discarded(card):
        return 'cardsdiscarded'

hand = ['C','C','D','X','D','E','C']

print Counter(
    x for x in (chamber_or_discard(card) for card in hand) if x is not None
)

Personally, I'd go for the class approach, perhaps even wrapping Counter, as it keeps all the associated functionality together.

Comments

[Matthew Trevor]

The recommendation from the docs is to never modify locals(), as the behaviour you're noting is entirely based on the implementation and is not guaranteed: "The contents of this dictionary should not be modified; changes may not affect the values of local and free variables used by the interpreter."

[toonarmycaptain]

chambersinreactor, cardsdiscarded = find_chamber_discard(chambersinreactor, cardsdiscarded) - this is amazing - who knew you could return multiple values like this?!