Local and global variables

The scope of a variable refers to the sections of a program where a variable is accessible. A local variable is only accessible in a defined section of the program, while a global variable is available for use in any section of the program.

Local variables

Variables defined within a Python function are local variables, only available within the function. This applies to both function parameters, and other variables defined within the function definition. A variable which is local to a function does not exist outside the function.

In the following example we are trying to access the variable x in the main function, but this causes an error:

def testing():
    x = 5
    print(x)

testing()
print(x)

The variable x only exists while the function testing is being executed. Other functions or the main function cannot access the variable.

Global variables

Variables defined within the main function are global variables. We previously defined the main function as those sections of code in a Python program which do not fall within any other function. The value stored in a global variable can be accessed from any other function in the program, so the following does not cause any errors:

def testing():
    print(x)

x = 3
testing()

A global variable cannot be changed directly from within another function. The following function has no effect on the value stored in the global variable:

def testing():
    x = 5
    print(x)

x = 3
testing()
print(x)

Here the function testing creates a new, local variable x, which "masks" the global variable while the function is being executed. This variable has the value 5, but it is a different variable than the global x which is defined in the main function.

But what would the following code do?

def testing():
    print(x)
    x = 5

x = 3
testing()
print(x)

The function testing assigns a value to the variable x, so Python interprets x to be a local variable instead of the global variable of the same name. The function attemts to access the variable before it is defined, so there is an error.

If we wish to specify that we mean to change the global variable within a function, we will need the Python keyword global:

def testing():
    global x
    x = 3
    print(x)

x = 5
testing()
print(x)

Now the assignment x = 3 within the function also affects the main function. All sections of the program are using the same global variable x.

When should you use global variables?

Global variables are not a way to bypass function parameters or return values, and they should not be used as such. That is, it is possible to write a function which stores its results directly in a global variable:

def calculate_sum(a, b):
    global result
    result = a + b

calculate_sum(2, 3)
print(result)

It is better to write a function with a return value, as we are used to do by now:

def calculate_sum(a, b):
    return a + b

result = calculate_sum(2, 3)
print(result)

The advantage of the latter approach is that the function is an independent whole. It has certain, defined parameters, and it returns a result. It has no side effects, so it can be tested and changed independently of the other sections of the program.

Global variables are useful in situations where we need to have some common, "higher level" information available to all functions in the program. The following is an example of just such a situation:

def calculate_sum(a, b):
    global count
    count += 1
    return a + b

def calculate_difference(a, b):
    global count
    count += 1
    return a - b


count = 0
print(calculate_sum(2, 3))
print(calculate_sum(5, 5))
print(calculate_difference(5, 2))
print(calculate_sum(1, 0))
print("There were", count, "function calls")

In this case we want to keep track of how many times either of the functions were called during the execution of the program. The global variable count is useful here, because we can increment it from the functions themselves as they are executed, but still access the final value in the main function.

Passing data from one function to another, revisited

If a program consists of multiple functions, the question of passing data from one function to another often comes up.

When we touched upon this topic previously, we had a program which asks the user for some integer values, prints them out, and performs some statistical analysis on the numbers. The program was divided into three separate functions:

def input_from_user(how_many: int):
    print(f"Please type in {how_many} numbers:")
    numbers = []

    for i in range(how_many):
        number = int(input(f"Number {i+1}: "))
        numbers.append(number)

    return numbers

def print_result(numbers: list):
    print("The numbers are: ")
    for number in numbers:
        print(number)

def analyze(numbers: list):
    mean = sum(numbers) / len(numbers)
    return f"There are altogether {len(numbers)} numbers, the mean is {mean}, the smallest is {min(numbers)} and the greatest is {max(numbers)}"

# the main function using these functions
inputs = input_from_user(5)
print_result(inputs)
analysis_result = analyze(inputs)
print(analysis_result)

An example of the program's execution:

The basic principle here is that the main function "stores" the data processed by the program. Here, this means the numbers typed in by the user, which are stored in the variable inputs.

If the numbers are needed in some function, the variable is passed as an argument, as seen above when the functions print_result and analyze are called. If the function produces a result that is relevant elsewhere in the program, the function returns this with a return statement, as seen with the functions input_from_user and analyse above.

As always in programming, there are many ways to arrive at the same functionality. It would be possible to use the keyword global and have the functions directly access the inputs variable defined in the main function. There are good reasons why this is not a smart move, however. If many different functions can access and potentially change a variable directly, it quickly becomes impossible to reliably track the state of the program, and the program risks becoming unpredictable. This is especially the case as the number of function involved grows large, as it is wont to do in large software projects.

In conclusion, passing data into and out of functions is best handled by arguments and return values.

You could also separate the implicit main function in the example above into its own, explicit main function. Then the variable inputs would no longer be a global variable, but instead a local variable within the main function:

# your main function goes here
def main():
    inputs = input_from_user(5)
    print_result(inputs)
    analysis_result = analyze(inputs)

    print(analysis_result)

# run the main function
main()

Please respond to a quick questionnaire on this week's materials.

You can check your current points from the blue blob in the bottom-right corner of the page.