Hanukkah of Data using Django

April-2025

Introduction

Using the Django ORM for solve the Hanukkah of Data Challenge turned out to be a great way to practice Django queries. You can view my repo on GitHub

Setting up Initial Database

We want to use the existing Sqlite file in the data folder.

The Sqlite file does not contain any constraints or indices. It may have been generated from the CSV by an automated tool. We will see that to have performant queries, we will need to improve this.

Django needs to initialise its ORM, where we can use the inspectdb flag to determine the shape of data

python manage.py inspectdb > models.py

By default it will have a very raw representation e.g.

class Customer(models.Model):
  customerid = models.IntegerField(blank=True, null=True)
  name = models.TextField(blank=True, null=True)
  address = models.TextField(blank=True, null=True)
  birthdate = models.TextField(blank=True, null=True)

We can do a basic amount of cleanup

Remove blank / null options where we know data exists
Set
```
customerid
```
as the primary key to avoid an automatic
```
id
```
column (note the original DB does not have a PK, but Django requires it)
Use a CharField as we know the length of the fields

class Customer(models.Model):
  customerid = models.IntegerField(primary_key=True)
  name = models.CharField()
  address = models.CharField()
  birthdate = models.CharField()

inspectdb creates unmanaged models by default docs That is, managed = False in the model’s Meta class. This tells Django not to manage each table’s creation, modification, and deletion. We will want to modify the tables in future, so we remove this line.

Now we need to generate a migration but we don't want to delete the existing database. We can use the flag --fake-initial to skip new table creations. Note now the schema will not exactly match.

python manage.py makemigrations
python manage.py migrate --fake-initial