Defining and calling functions

A function is a separate piece of code that can be used multiple times in different places in the program. So far, the use of functions, such as print, len, range, etc. has been shown several times. You did not have to write it yourself, because it was already done by the creators of Python. It was enough to know how they work. Most functions require external information - numbers, texts, other objects - and they are given as arguments to the function. Most also return some result, which, for example, can be assigned to a variable or used directly in an expression.

Defining a function

In every modern programming language it is possible to define functions independently. Without them, teamwork on even an average program would be almost impossible. New functions should be created whenever we are able to extract a code fragment that performs a certain, defined operation for certain data (and possibly returns a result). Also, when writing a program, you see the need to use an identical piece of code several times, you should put it in a separate function - thanks to this, the code only needs to be written once and you can easily use it multiple times.

We define functions using the def keyword:

def function_name (argument1, argument2 ...):
block containing commands
performed in a function

The name of the function and the names of the arguments must follow exactly the same rules as the names of the variables. For example, below is the definition of a function that prints a string in a frame:

def print_int_frame (text):
    line = "+-" + "-" * len(text) + "-+"
    print(line)
    print("|", text, "|")
    print(line)

Once a function has been defined, it is called with its name followed by parentheses containing arguments:

print_int_frame("Hello!")

print_int_frame("How are you?")

Note! Even if the function takes no arguments, calling it requires giving the parentheses, even if nothing is in them:

def greet ():
print ("How are you?")

greet()

Returning a value

Functions can also return values as the result of their operation. The return keyword is used for this, which can only appear inside a function. After this keyword, we put the value to be returned. For example

def sum(number1, number2):
    return number1 + number2

When calling such a function, the result returned by it can be used in any way, e.g. by assigning it to a variable or using it directly in an expression (it has already been demonstrated many times, e.g. using the len or range functions). In the case of the above function, it could be:

number1 = float(input("Enter the first number:"))
number2 = float(input("Enter the second number:"))

print("The sum of these numbers is:", sum(number1, number2))

Within a function, the return command can be anywhere, including a conditional block or inside a loop. In any case, it will cause the function to exit immediately: the execution of the loop will be interrupted and the commands below will not be executed. For example:

def quotient(dividend, divisor):
    if divisor == 0:
        return
    quotient = dividend / divisor
    return quotient

The return command itself, with no return value given, will return None. The same happens when the function ends without the return command (as in the first example, the print_in_frame function). So, in Python, each function returns its result: defaulting to None.

The function can also return several results simultaneously. To do this, all results should be given after the return command, separated by a comma. For example

def sum_and_difference(number1, number2):
    sum = number1 + number2
    difference = number1 - number2
    return sum, difference

When calling such a function, its result can be assigned to two variables, separated by a comma:

sum21, difference21 = sum_and_difference (2, 1)

You can also assign the results of such a function to one variable or use it directly in an expression. In this case, it will be treated as a tuple containing a given number of elements:

result = sum_and_difference(5, 3)  # result = (8, 2)
print(type(result))                # tuple

The type function used in the above example returns the type of the value given as its argument.

Local and global variables

Variables defined outside of functions (as in all the previous lessons) are global variables. They are available throughout the program from the moment of their definition (value assignment). In some of the examples above, variables were also defined inside a function. These are local variables and they are shown exclusively inside the function, and after leaving they are destroyed. Function arguments are also treated as local variables (with values defined at the time the function was called).

If there are variables with the same names inside and outside the function, the values assigned to the local variable will not change the global variable:

def function (): 
    variable = 2
    print ("Local variable:", variable )     # 2

variable = 1
function()
print ("Global variable:", variable )       # 1

The two variables in the example above have the same name. However, they are completely independent of each other. There is no access to the global variable inside the function because it is “obscured” by a local variable with the same name. If we do not assign any value to the variable inside the function, then we can read and use the value of the global variable:

def function():
    print ("Global variable:", variable ) # 1

variable = 1
function()

Attention! For a variable to be treated as a local variable, it is enough that it is assigned a value anywhere inside the function. The following code won’t work:

def function ():
    print ("Global variable:", variable ) # ERROR - local variable is not defined yet
    variable = 2

variable = 1
function()

It is possible to force the variables inside the function to be treated as global by a command global followed by a list of global variable names. Assigning values to them inside this function changes the global variable:

def function():
    global variable
    variable = 2
    print ("Global variable:", variable ) # 2

variable = 1
function()
print ("Global variable:", variable ) # 2

An important rule!

Each extractable code fragment should be placed in a separate function. Also, make sure that functions that perform calculations do not print anything on the screen or change any global variables (generally avoid using a keyword global). When writing a function, we are often unable to predict how many times it will be performed. These can be thousands of times - in which case any message inside the function will be printed thousands of times. In general, parts of the program responsible for calculations and for interaction with the user (e.g. through functions print and input) should be clearly separated from each other. Thanks to this, in the future it will be possible and relatively simple to change the way of communication with the user, e.g. to a windowed version or via a website. Under no circumstances you may use input in functions which purpose is different than directly reading from the keyboard All necessary data must be given as arguments to the function and the result returned with return.

Calling a function

The function is always called by giving its name followed by parentheses. The arguments of the function are placed inside these parentheses. Their number must be the same as the number of arguments given when defining the function (except for default arguments and functions with a variable number of arguments, which are described below).

There are two ways to pass arguments to a function: by position and by name. To explain this, suppose we have a defined function:

def function(a, b, c, d, e):
   function block

It takes 5 arguments. In the basic method of calling a function, we give 5 values in parentheses, separated from each other by commas:

function(1, 2, 3, 4, 5)

Then the first argument (a) will be 1 for a given call, the second (b) will be 2, etc. When calling a function, its arguments can also be passed by name by giving argument_name=value (with a single = character as in the variable assignment command):

function(a=1, c=3, b=2, e=5, d=4)

In this case, the order in which we gave the arguments does not matter. The argument a will be 1, b: 2, c: 3, d: 4, and e: 5. It is also possible to combine both of the above methods. Then we first pass any number of arguments by position, then arguments by name (no need to follow the order or arguments passed by name):

function(1, 2, d=4, c=3, e=5)

In this case, the values 1 and 2 will be assigned to the arguments a and b (as they appear in the definition as the first and second arguments), and the remaining arguments are passed by name (that is c: 3, d: 4, and e: 5).

Note! Once we start supplying arguments using their names, we must not revert to positional specification on the same call. This will cause a syntax error.

Default arguments

It often happens that some arguments of a function have sensible defaults. Then, when defining the function, we can give these default values using the sign =:

def function(a, b, c, d=0, e=None):
   function block

The important thing is that all arguments with default values must be at the end of the argument list . This means that if a default value is given for any argument, each subsequent argument must also have a default value.

When calling a function defined in this way, we can omit these arguments. Then, when invoked, their default values will be adopted:

function(1, 2, 3)

In the above call, the value of the argument d will be 0, and the argument e: None. Calling:

function(1, 2, 3, 4)

In this call, the argument d will be set to 4 and e will remain at its default value.

When using default arguments, it is very useful to pass arguments by name. Let’s see the following example of a call:

function(1, 2, 3, e=5)

In this case, the argument d will keep its default value, and the argument e will take the value given in the function call (5). Python often uses functions with a very large number of arguments, most of which have default values. When calling such a function, you can easily change only the arguments you need.

Attention! Default argument values are evaluated when the function is defined, not when it is invoked. For this reason, you should not use mutable types (lists and dictionaries) as default arguments. Instead of

def function(numbers=[1, 2, 3]): 
    function body

you should write

def function(numbers=None): 
    if numbers is None: 
        numbers = [1, 2, 3]
    function body

Recurrence

A function can call itself to solve a problem. We call such a function a recursive function. Example calculating the factorial:

def factorial(n):
     if n == 1:
         return 1
     elif n > 1:
         return n * factorial (n-1) 
     else:
         return None

For a recursive function to work properly, each subsequent call must be made for a different value of the argument controlling the the recurrence (in the example above it is factorial(n-1)). It must also be guaranteed to stop subsequent function calls when the value of this argument reaches a given value (in the example it will be for n equal to 1).

Published under Creative Commons Attribution-NonCommercial-ShareAlike license.