Python Data Class: A Better Way to Store Data

A Python data class is a regular Python class that has the @dataclass decorator. It is specifically created to hold data. Since Python version 3.7, Python offers data classes through a built-in module that you can import, called dataclass. There are several advantages over regular Python classes which we’ll explore in this article. We’ll also look at example code and a couple of common operations you might want to perform with data classes.

The advantage of using data classes

Why should you use a data class instead of a regular Python class? First, let’s look at some of the advantages a Python data class has to offer.

Requires a minimal amount of code

The @dataclass decorator adds a lot of functionality to a class without adding any visible code. This allows your data class to be very compact while still offering many useful features. All you need to do is define the fields to hold your data. You don’t need to define any functions.

Comparison

Two Python data classes can be compared with == because the so-called dunder method __eq__ is implemented automatically. In general, we can compare any Python object that implements this special method to other objects of the same type.

Printing a data class

Similarly, because __repr__ is implemented, you can print data classes and get a nice representation of it. This is especially useful for debugging.

Data classes require type hints

Data classes are built around the new(ish) type system Python offers. Using type hints reduces the chances of bugs and unexpected behavior in your code. You essentially declare the type of data that should be stored in a variable.

Python data class example

Here’s an example of a data class at work:

from dataclasses import dataclass

@dataclass
class Card:
    rank: str
    suit: str
    
card1 = Card("Q", "hearts")
card2 = Card("Q", "hearts")

print(card1 == card2)
# True

print(card1.rank)
# 'Q'

print(card1)
Card(rank='Q', suit='hearts')

Default values

A data class can have default values. Assigning default values is as simple as assigning a value to a variable. For example, to make our Card class have a default value of Queen of hearts, we can do as follows:

from dataclasses import dataclass

@dataclass
class Card:
    rank: str = 'Q'
    suit: str = 'hearts'

Converting a data class to JSON

A common use case is to convert your nicely structured data class to JSON. E.g., if you want to export the data to a database, or send it to the browser. The bad news here: there’s no built-in way to convert a data class to JSON. At least not in such a way that it can comfortably export all kinds of data types inside your class (like date objects).

The good news is, that there is a Python package called dataclasses-json that simplifies the task. However, it requires an extra decorator. You’ll need to install the package with the pip install command or something like Pipenv, preferably inside a virtual environment. For example:

$ pip install dataclasses-json

Here’s an example of how you can use the package:

from dataclasses import dataclass
from dataclasses_json import dataclass_json

@dataclass_json
@dataclass
class Card:
    rank: str = 'Q'
    suit: str = 'hearts'

card = Card()
print(card.to_json())

This results in the following output:

{"rank": "Q", "suit": "hearts"}

Another method is to use Python inheritance and inherit from the JSONEncoder class to create your own custom encoder. The advantage here is that you don’t need to install an external package. You can learn how to do this in this blog post.

Keep learning

About Erik van Baaren

Erik is the owner of Python Land and the author of many of the articles and tutorials on this website. He's been working as a professional software developer for 25 years, and he holds a Master of Science degree in computer science. His favorite language of choice: Python! Writing good articles takes time and effort. Did you like this tutorial? You can buy him a coffee to show your appreciation.

Subscribe
Notify of
1 Comment
Oldest
Newest Most Voted
Inline Feedbacks
View all comments

No, thanks!