In my 8 years of working on large Python codebases, I think the Django ORM is one of the best pieces of software I've had the joy of using. I often recommend it independent of the rest of Django (e.g. to people building Flask apps that need an ORM+migrations).
Especially when working on a team with many junior developers who don't know SQL (myself included in the early days), the Django ORM has been a shining beacon of reliability and consistency allowing us to do complex database schema changes without ever having to worry about messing them up or not being able to revert our changes.
I don't think I can say the same for any other ORM I've worked with, in particular I have really bad memories of trying to match Django's migration reliability and flexibility using SQLAlchemy + Alembic.
Also in regards to switching databases, the ability to swap out PostgreSQL for SQLite on smaller projects without having anything break is a feature I've relied on several times, and I'm extremely grateful maintaining that clean break in the layers of abstraction was prioritized by the Django team.
Obviously the author of this article has had different experiences (and has considerable authority in this field), but I caution that their experiences are not universal among Python devs.
Especially when working on a team with many junior developers who don't know SQL (myself included in the early days), the Django ORM has been a shining beacon of reliability and consistency allowing us to do complex database schema changes without ever having to worry about messing them up or not being able to revert our changes.
I don't think I can say the same for any other ORM I've worked with, in particular I have really bad memories of trying to match Django's migration reliability and flexibility using SQLAlchemy + Alembic.
Also in regards to switching databases, the ability to swap out PostgreSQL for SQLite on smaller projects without having anything break is a feature I've relied on several times, and I'm extremely grateful maintaining that clean break in the layers of abstraction was prioritized by the Django team.
Obviously the author of this article has had different experiences (and has considerable authority in this field), but I caution that their experiences are not universal among Python devs.