Functions in programming allow reusing and logically organising programs into blocks of code or modules. We have already used print, len, hasattr, and isinstance, which are all examples of built-in Python functions.

We have also used methods, str, int, float, and type. Method is a concept from object-oriented programming that refers to “special” functions that belong to specific objects or classes (of objects).

In Python everything is an object (e.g., str, int, and float objects), including the functions. The functions have their own type, function.

We can’t avoid using objects in Python. That said, object-oriented programming is optional in Python.

Defining Functions

Python uses def keyword to define a function. You will need to provide a name, and optionally input parameters (a.k.a arguments). A function also must have a body, an indented block of code that runs each time you call the function.

def FUNCTION_NAME(INPUT_ARGS):
    BODY
    return RESULTS

To define an empty (template) function, that you plan to implement later, use pass keyword. You must always include () and : after the name of the function, even when you don’t need to provide any input.

def FUNCTION_NAME():
    pass

It is conventional to use four spaces or one tab for each level of indentation. Many text editors and IDEs allow setting dafault identation. Be consistent with the size and type (spaces or tabs) of indentation you use throughout your code. This is mainly for readability but also to avoid bugs due to inconsistent indentation.

Return Statement

The optional return statement stops the execution and returns the results of the function’s computations up to that point.

  • If we don’t provide any variables to return, the function stops at the return statement and returns None.
  • If we completely omit the return statement, entire function body is executed, and once completed, the function returns None.

Multiparameter Functions

For many applications you would need more than one input parameter. E.g., a function add(a, b) with two inputs a and b. To define a function that accepts multiple input parameters, simply list the inputs in the function definition as a comma separated sequence.

def func(input1, input2, input3):
    pass

Returning Multiple Values

Similarly, a function also can return multiple values by specifying multiple variables in the return statement.

def func():
    ...
    return result1, result2, ...

args and kwargs

For the positional arguments, the value is defined by its position in the function definition. E.g., all the functions above used positional parameters. Python also allows defining default values for function arguments as key-value pairs. Key in this case refers to the variable name that your function uses inside its body.

def func(arg1, arg2, value = 10.5, status = "active", on = True):
    ...

In the above, arg1 and arg2 are positional arguments (args) and the rest are keyword arguments (kwargs) with default values.

In Python documentation, positional arguments are commonly called args and keyword arguments are called as kwargs.

  • The argument’s value is inferred from its position if we don’t provide the key for the argument.
new_T = decrease_val(25, 10)
  • You can also specify the name of the positional argument to provide values with their names inside the function.
new_T = decrease_value(delta=10, current=25)

Positional arguments must be provided before keyword arguments when defining and calling functions that use both positional and keyword arguments.

Exercises

E1: Fix a Greeter Function.

  • Run the code. It should fail to run.
  • Fix the code so that it prints Hello, World!.
  • Set the name variable to your name as a str (using quotes or double quotes) then run the code again to print a geeting.

E2: Complete the vector_length function.

  • Compute the Euclidean length or L2-norm of a 3-dimensional vector by completing the function.
  • You may find sqrt function useful for this. I have already imported it for you as a part of built-in math module. You can call it with math.sqrt(...) as shown below.

Modules are files with reusable code. E.g., you can write your own module as text file ending with .py with your functions and variables (a.k.a. Python file) and then import and reuse it in other parts of your code.

E3: Complete function compute_l1_l2 below.

  • The function should return two values. First value must be the Manhattan length l1 and the second returned value must be the Euclidean length. Use predefined functions to compute L1 (Manhattan) and L2 (Euclidean) norms.
  • You are including function calls for l1_norm and l2_norm before Python is aware of them. Remembering that Python executes code sequentially, why doesn’t this lead to an error? Tip: function executes (runs) only when you call it.
  • What is the sequence of calls when you run the code above?

E4: Function Objects

As noted earlier on this page, Python treats functions as objects. You can check this using type method. Keeping this in mind, answer following question by running the code cell below.

  • What is the difference between hi and hi() below? Print statements are there to examine these objects by printing them to the console.

You don’t need to use print to examine these commands in the Python shell.