Dunder/Magic Methods in Python

March 23, 2021

Dunder methods are names that are preceded and succeeded by double underscores, hence the name dunder. They are also called magic methods and can help override functionality for built-in functions for custom classes.

1. Introduction

Implementing dunder methods for classes is a good form of Polymorphism. If you have ever created a class in Python and used the init function, then you have already been using dunder methods.

2. Table of contents

3. Prerequisites

Before we continue it will be important to have the following:

  • A basic understanding of Object Oriented Programming using Python.
  • Experience working with classes in Python.
  • Familiarity with built-in functions such as len, get, set, etc.

4. Why do we need Dunder methods?

Consider a case where we have the following class:

class point:
    x = 4
    y = 5

p1 = point()
print(p1)

The print statement would print something like <__main__.point object at 0x7fb992998d00>. But, we might want the print statement to display something in the format (4,10). We can achieve this by overriding the __str__ method of our class.

We could also override other methods such as the len, +, [] etc. We will create a new class and override many of the built-in functions in this article.

5. Our custom class

class softwares:
    names = []
    versions = {}

This class will be used to save a list of software and their versions. names is a list to store the names of the software and versions is a dictionary where the key is the software name and the value is the version number. By default, all softwares start with a version of 1.

6. Dunder methods for our class

Before moving on, please ensure that your indentation is correct. The methods that will be discussed below are methods belonging to the class we created and must be indented appropriately.

6.1. init

This is a method you must have already used if you have worked with classes. The init method is used to create an instance of the class.

def __init__(self,names):
    if names:
        self.names = names.copy()
        for name in names:
            self.versions[name] = 1
    else:
        raise Exception("Please Enter the names")

The init method defined above accepts a list of names as parameters and stores it in the class’ names list. Additionally, it also populates the versions dictionary. We have also put a check on the names list.

If the list is empty, an exception is raised. Below is how we would use the init method.

p = softwares(['S1','S2','S3'])
p1 = softwares([])

The first statement would work fine but the second line would raise an exception since an empty list was passed in as a parameter.

6.2. str

The str method is useful when we want to use instances of our class in a print statement. As discussed earlier, it usually returns a memory object. But we can override the str method to meet our requirements.

def __str__(self):
    s ="The current softwares and their versions are listed below: \n"
    for key,value in self.versions.items():
        s+= f"{key} : v{value} \n"
    return s

The above str method returns the software and their versions. Ensure that the function returns a string. Below is how we would call the method.

print(p)

6.3. setitem

When assigning values in a dictionary, the setitem method is invoked.

d = {}
d['key'] = value

We can give instances of our class a similar feature with the help of the setitem method.

def __setitem__(self,name,version):
    if name in self.versions:
        self.versions[name] = version
    else:
        raise Exception("Software Name doesn't exist")

The method above is going to update the version number of the software. If the software is not found, it will raise an error.

In the 3rd line, we use the built-in setitem method of a dictionary.

We can invoke the setitem method in the following way:

p['S1'] = 2
p['2'] = 2

The first line would update the version of software S1 to 2. But the second line would raise an exception since software 2 doesn’t exist.

6.4. getitem

The getitem method is like the setitem method, the major difference being that the getitem method is called when we use the [] operator of a dictionary.

d = {'val':key}
print(d['val'])

Instances of our class can also be given a similar feature.

def __getitem__(self,name):
    if name in self.versions:
        return self.versions[name]
    else:
        raise Exception("Software Name doesn't exist")

The above method essentially returns the version of the software. If the software is not found, it raises an exception. To invoke the getitem method, we can write the following line of code.

print(p['S1'])
print(p['1'])

The first line would print the version of S1. But, the second line would raise an Exception since 1 doesn’t exist.

6.5. delitem

The delitem is like the setitem and getitem method. To avoid repetition, we will move on to the implementation and use case.

def __delitem__(self,name):
    if name in self.versions:
        del self.versions[name]
        self.names.remove(name)
    else:
        raise Exception("Software Name doesn't exist")

The delitem method deletes the software from the dictionary as well as the list.

It can be used as follows.

del p['S1']

6.6. len

In a dictionary, the len method returns the number of elements in a list or the number of key-value pairs in a dictionary.

We can define a len method for our class as well.

def __len__(self):
    return len(self.names)

The len method for our class returns the number of softwares. As you might have noticed, we are using the built-in len method of a list to return the number of software.

The len method of our class can be used in the following way.

print(len(p))

6.7. contains

The contains method is used when using the in operator. The return value has to be a boolean.

def __contains__(self,name):
    if name in self.versions:
        return True
    else:
        return False

The method checks if the name is found in the dictionary. We will be using the dictionary’s built-in contains method for that.

if 'S2' in p:
    print("Software Exists")
else:
    print("Software DOESN'T exist")

The code above prints the statement inside the if blocks since software S2 is present inside the versions dictionary.

6.8. Complete code

class softwares:
    names = []
    versions = {}
    
    def __init__(self,names):
        if names:
            self.names = names.copy()
            for name in names:
                self.versions[name] = 1
        else:
            raise Exception("Please Enter the names")
    
    def __str__(self):
        s ="The current softwares and their versions are listed below: \n"
        for key,value in self.versions.items():
            s+= f"{key} : v{value} \n"
        return s
    
    def __setitem__(self,name,version):
        if name in self.versions:
            self.versions[name] = version
        else:
            raise Exception("Software Name doesn't exist")
    
    def __getitem__(self,name):
        if name in self.versions:
            return self.versions[name]
        else:
            raise Exception("Software Name doesn't exist")
    
    def __delitem__(self,name):
        if name in self.versions:
            del self.versions[name]
            self.names.remove(name)
        else:
            raise Exception("Software Name doesn't exist")
    
    def __len__(self):
        return len(self.names)
    
    def __contains__(self,name):
        if name in self.versions:
            return True
        else:
            return False

7. Some more dunder methods

Before looking at some more dunder methods, let’s create a new class.

class point:
    x = None
    y = None
    
    def __init__(self, x , y):
        self.x = x
        self.y = y
    
    def __str__(self):
        s = f'({self.x},{self.y})'
        return s

p1 = point(5,4)
p2 = point(2,3)

We have created a class point which is basically a 2D point. The class has an init method and a str method. We have also created a couple of instances of the class.

7.1. add

The add method is called when using the + operator. We can define a custom add method for our class.

p1 + p2 is equal to p1._add__(p2)

def __add__(self,p2):
    x = self.x + p2.x
    y = self.y + p2.y
    return point(x,y)

The above method adds the x and y coordinates of the first instance of point and the second instance of point. It will create a new instance of point and then return it.

p3 = p1 + p2

The line of code above invokes the add method.

7.2. iadd

The iadd method is like the add method. It is invoked when using the += operator

def __iadd__(self,p2):
    self.x += p2.x
    self.y += p2.y
    return self

The method above just updates an instance’s coordinates by adding the coordinates of p2. Make sure you are returning self, otherwise it will return None and won’t work as expected.

p1 += p2
print(p1)

The above method invokes the iadd method.

7.3. Other operators

  • __sub__(self,p2) ( - )
  • __isub__(self,p2) ( -= )
  • __mul__(self,p2) ( * )
  • __imul__(self,p2) ( *= )
  • __truediv__(self,p2)( \ )
  • __itruediv__(self,p2) ( \= )
  • __floordiv__(self,p2) ( \\ )
  • __ifloordiv__(self,p2) ( \= )

7.4. call

When invoking a function like func(), we are invoking the call method.

If we put in place a call method for our class, we can do the following:

p1()
p2()

Below is an example call method:

def __call__(self):
    print(f"Called Point {self.x},{self.y}")

7.5. Complete code

class point:
    x = None
    y = None
    
    def __init__(self, x , y):
        self.x = x
        self.y = y
    
    def __str__(self):
        s = f'({self.x},{self.y})'
        return s
    
    def __add__(self,p2):
        print("In add")
        x = self.x + p2.x
        y = self.y + p2.y
        return point(x,y)
    
    def __iadd__(self,p2):
        self.x += p2.x
        self.y += p2.y
        return self
    
    def __isub__(self,p2):
        self.x -= p2.x
        self.y -= p2.y
        return self
    
    def __imul__(self,p2):
        self.x *= p2.x
        self.y *= p2.y
        return self
    
    def __itruediv__(self,p2):
        self.x /= p2.x
        self.y /= p2.y
        return self
    
    def __ifloordiv__(self,p2):
        self.x //= p2.x
        self.y //= p2.y
        return self
    
    def __call__(self):
        print(f"Called Point {self.x},{self.y}")

8. Conclusion

Dunder methods are indeed magical and can help you improve the functionality of your class. You can find more dunder methods here.

Happy Learning! :)


Peer Review Contributions by: Peter Kayere


About the author

Rahul Banerjee

Rahul Banerjee is a Computer Engineering Undergrad student at the University of Toronto. He is currently doing an Internship at Ernst and Young. He is a Data Science enthusiast interested in writing tech articles on Medium. If he is not working or writing articles, he is usually binge-watching shows on Netflix.

This article was contributed by a student member of Section's Engineering Education Program. Please report any errors or innaccuracies to enged@section.io.