Case insensitive unique model fields in Django?

With overriding the model manager, you have two options. First is to just create a new lookup method:

class MyModelManager(models.Manager):
   def get_by_username(self, username):
       return self.get(username__iexact=username)

class MyModel(models.Model):
   ...
   objects = MyModelManager()

Then, you use get_by_username('blah') instead of get(username='blah'), and you don't have to worry about forgetting iexact. Of course that then requires that you remember to use get_by_username.

The second option is much hackier and convoluted. I'm hesitant to even suggest it, but for completeness sake, I will: override filter and get such that if you forget iexact when querying by username, it will add it for you.

class MyModelManager(models.Manager):
    def filter(self, **kwargs):
        if 'username' in kwargs:
            kwargs['username__iexact'] = kwargs['username']
            del kwargs['username']
        return super(MyModelManager, self).filter(**kwargs)

    def get(self, **kwargs):
        if 'username' in kwargs:
            kwargs['username__iexact'] = kwargs['username']
            del kwargs['username']
        return super(MyModelManager, self).get(**kwargs)

class MyModel(models.Model):
   ...
   objects = MyModelManager()

Store the original mixed-case string in a plain text column. Use the data type text or varchar without length modifier rather than varchar(n). They are essentially the same, but with varchar(n) you have to set an arbitrary length limit, that can be a pain if you want to change later. Read more about that in the manual or in this related answer by Peter Eisentraut @serverfault.SE.

Create a functional unique index on lower(string). That's the major point here:

CREATE UNIQUE INDEX my_idx ON mytbl(lower(name));

If you try to INSERT a mixed case name that's already there in lower case you get a unique key violation error.
For fast equality searches use a query like this:

SELECT * FROM mytbl WHERE lower(name) = 'foo' --'foo' is lower case, of course.

Use the same expression you have in the index (so the query planner recognizes the compatibility) and this will be very fast.


As an aside: you may want to upgrade to a more recent version of PostgreSQL. There have been lots of important fixes since 8.4.2. More on the official Postgres versioning site.


As of Django 1.11, you can use CITextField, a Postgres-specific Field for case-insensitive text backed by the citext type.

from django.db import models
from django.contrib.postgres.fields import CITextField

class Something(models.Model):
    foo = CITextField()

Django also provides CIEmailField and CICharField, which are case-insensitive versions of EmailField and CharField.


As of December 2021, with the help of Django 4.0 UniqueConstraint expressions you can add a Meta class to your model like this:

class Meta:
    constraints = [
        models.UniqueConstraint(
            Lower('<field name>'),
            name='<constraint name>'
        ),
    ]

I'm by no mean a Django professional developer and I don't know technical considerations like performance issues about this solution. Hope others comment on that.