Django-nonrel vs Django with PyMongo/Ming for use with MongoDB

Question

I'm about to start a project with Django and MongoDB. From what I can tell, there are two ways to use Django and MongoDB together. One way is the fork of Django called Django-nonrel which supports NoSQL databases, and the other is to just use the standard version of Django with PyMongo (and optionally Ming for models, etc.).

I have used the second option lightly in the past, but this is going to be a large project. I honestly don't know too much about Django-nonrel, but it seems like using the primary Django project would be better from a stability standpoint, as well as from a new features standpoint.

Is there an advantage to using Django-nonrel over just base Django with PyMongo thrown in?

Answer 1

This is a biased opinion, but I would use Django with PyMongo. The main reasons are the following. First, what you stated about using the standard Django version versus a fork. And secondly, the fact that PyMongo is the official Python driver for MongoDB. It is written and maintained by MongoDB. It ensures that any new feature in MongoDB will be available in PyMongo.