Getting Started with Django Polymorphic Models

Polymorphism is a technique for modeling data in a database. It allows applications to store different types of data in the same database table. <!--more--> This tutorial will model an online book store dealing with print books and ebooks sales.

We will begin by creating simple models and adding more fields to our models as the online store expands and handles more products.

In an online bookstore, a user can order an ebook version of a novel while another user orders a print version of the same book.

In the database, it's the same book but with different properties. The ebook version will include a download link which the print version won't have.

The print version can have a weight property that we can use to calculate the ebook version's delivery charges.

Table of contents #

• Table of contents
• Prerequisites
• Polymorphic modeling
  • Default modeling
  • Sparse modeling
  • Semi-structured modeling
  • Abstract base modeling
• Conclusion

Prerequisites #

To follow along with this article, you need to have:

• Python and virtualenv installed on your computer.

• Some knowledge of Django and Python.

Polymorphic modeling #

Django ORM provides several ways to model polymorphic data. As we will study below, we can either use the standard Django features or the advanced Django ORM features.

We will begin with the simplest method to the complex way of implementing polymorphic models.

The first step is to create a new Django project that we will use throughout this guide.

Create a new working directory named models by executing the following command:

mkdir modeling

Execute the command below to change the working directory to the one we have created above and create a new virtual environment.

Django projects use virtual environments to manage the project's dependencies in isolations to avoid dependency conflict in other projects.

virtualenv venv
source venv/bin/activate
pip install django

Execute the following command to create a new Django project named polymorphic:

django-admin startproject polymorphic

Since the Django project is organized into applications, execute the command below to create a modeling application. This is where we will model our application database:

python manage.py startapp modeling

Django provides several modeling patterns that we will go through in this guide. They include:

• Default modeling pattern
• Sparse modeling pattern
• Semi-structured modeling pattern
• Abstract modeling pattern

Default modeling #

This section will model our online bookstore using the default Django modeling features. Add the following code in the models.py file:

from django.contrib.auth import get_user_model
from django.db import models

# Publication is a journal that can be ordered from the bookstore a hard copy file
class Publication(models.Model):
    publication_title = models.CharField(max_length=50)
    publication_price = models.PositiveIntegerField()
    publication_weight = models.PositiveIntegerField()  # The weight of the print publication is required for calculations of the delivery fees

    def __str__(self) -> str:
        return self.publication_title


# UserCart manages the user publications add on the cart
class UserCart(models.Model):
    visitor_cart = models.OneToOneField(
        get_user_model(),  # Maps the visitor fields with the currently logged-in user
        primary_key=True,
        on_delete=models.CASCADE,
    )
    publications = models.ManyToManyField(Publication)

In the code snippet above, we create a publication model which will hold the information about the books being sold in the store.

The UserCart model holds the information about the currently logged-in user and the publications a user has added to the cart before checking out.

Advantages	Disadvantages
Easy to model and maintain	Only suitable for products with the same attributes

Sparse modeling #

Now that our online store has gained several customers, they request ebooks instead of print books. Therefore, we need to modify the Publication model to accommodate print and ebooks.

Replace the code snippet in the models.py file with the code below:

# Publication is a journal that can be ordered from the bookstore a hard copy file
class Publication(models.Model):
    PUBLICATION_CHOICES = (
        ("Print", 'Print'),
        ("Ebook", 'Ebook'),
    )
    publication_title = models.CharField(max_length=50)
    publication_price = models.PositiveIntegerField()
    publication_type = models.CharField(
        max_length=50,
        choices=PUBLICATION_CHOICES,
    ) # Publication type is required to check wether a download link is required for the publication
    publication_download_link = models.URLField(null=True,
                                                blank=True, )  # The ebook publications require a download link that can be used to download the book
    publication_weight = models.PositiveIntegerField()  # The weight of the print publication is required for calculations of the delivery fees

    def __str__(self) -> str:
        return self.publication_title


# UserCart manages the user publications add on the cart
class UserCart(models.Model):
    visitor_cart = models.OneToOneField(
        get_user_model(),  # Maps the visitor fields with the currently logged-in user
        primary_key=True,
        on_delete=models.CASCADE,
    )
    publications = models.ManyToManyField(Publication)

In the code snippet above, we added two fields to the Publication model.

The publication_weight property stores the weight of the print publication. We can use this field to calculate the delivery charges of the publication.

The publication_download_link field stores the download link for the publication if it is an ebook.

Advantages	Disadvantages
Easy to model and maintain	Does not utilize the NOT NULL database constraints
	Adding new attributes to the model requires model changes

Semi-structured modeling #

With the increase in book sales in our online store, there is an increase in the number of nullable fields. Unfortunately, this is becoming hard to maintain.

To address the increasing number of nullable fields, we can use a JSONField to store other properties and a model to store the common properties.

Update the models.py file in the modeling application with the code below:

from django.contrib.auth import get_user_model
from django.db import models

# Publication is a journal that buyers can order from the bookstore a hard copy file
from django.db.models import JSONField

class Publication(models.Model):
    PUBLICATION_CHOICES = (
        ("Print", 'Print'),
        ("Ebook", 'Ebook'),
    )
    publication_title = models.CharField(max_length=50)
    publication_price = models.PositiveIntegerField()
    publication_type = models.CharField(
        max_length=50,
        choices=PUBLICATION_CHOICES,
    )  # Publication type is required to check wether a download link is required for the publication
    publication_download_link = models.URLField(null=True,
                                                blank=True, )  # The ebook publications require a download link that can be used to download the book
    publication_extra_fields = JSONField() # The extra fields will be stored in the database as Json field

    def __str__(self) -> str:
        return self.publication_title

# UserCart manages the user publications add on the cart
class UserCart(models.Model):
    visitor_cart = models.OneToOneField(
        get_user_model(),  # Maps the visitor fields with the currently logged-in user
        primary_key=True,
        on_delete=models.CASCADE,
    )
    publications = models.ManyToManyField(Publication)

The semi-structured model allows us to keep the common fields on the model and other fields like weight and download link in the JSONField.

Advantages	Disadvantages
Reduces the number of nullable fields	Complex validation, we have to validate all the JSON data fields independently before saving
Easier to add new attributes.	Restricted database support since not all databases support JSON as a data type

Abstract base modeling #

So far, we have modeled our product using a single Django model. Our online book store has expanded, and we now want to sell other products i.e journals.

Our previous model design is not efficient enough to allow us to sell different types of products. Therefore, we need to develop an optimal solution for our database model.

Update the models.py in the modeling application with the code snippet below:

from django.db import models

# Publication is a journal that can be ordered from the bookstore a hard copy file
from django.db.models import JSONField

# This model represents the pdf downloadable version of a publication
class Ebook(models.Model):
    ebook_download_link = models.URLField()

# This the new type of publication that the users can download
class Journal(models.Model):
    journal_download_link = models.URLField()

• In the code snippet above, we create a standard model Publication that other products will extend. With this type of design, the common properties of the products we sell are in the Publication model, while the specific properties of the product we sell are in the particular product model. For example, the Ebook product has a specific field book_weight.

Advantages	Disadvantages
Easy to maintain, design, and test	Hard to scale since every new product requires an additional model

Conclusion #

In this article, you have learned how to create polymorphic Django models.

You can now implement polymorphic Django models in your application to reduce complexity. Ensure that you use a suitable polymorphic approach that fits your use case.

---

Peer Review Contributions by: Odhiambo Paul