`super()` and `__init__()` in Python

David Y.

The Problem

What does super() refer to in Python and why is it often used in __init__() methods?

The Solution

In Python, super() is a built-in function used to call methods defined in the parent class. One of the advantages of inheritance in an object-oriented language like Python is avoiding code duplication. Consider the following code:

MARKUP = 1.5

class Product:
    def __init__(self, name, cost_price):
        self.name = name
        self.cost_price = cost_price
        self.selling_price = cost_price * MARKUP

class Perishable(Product):
    def __init__(self, name, cost_price, shelf_life):
        super().__init__(name, cost_price)
        self.shelf_life = shelf_life

The line starting with super().__init__ will execute the __init__ method of Product as if it were a method defined in Perishable. This saves us from having to rewrite our pricing code in every child class of Product. Whenever we create an instance of Perishable, it will have a name, cost_price, selling_price, and shelf_life.

In this example, we could replace super().__init__ with Product.__init__ and achieve the same thing, but super() allows us to write more general code and is necessary for more complex class hierarchies that use dependency injection and multiple inheritance.

In Python 2 code and older Python 3 code, you may encounter lines written like this:

super(Perishable, self).__init__(name, cost_price)

This is equivalent to calling super() in Python 3. While it is more explicit and was the required way to use super() in Python 2, it is no longer necessary, will slow down our code, and should thus be avoided.