django-sabridge - SQLAlchemy access to Django models



To demonstrate sabridge, we will access django.contrib.auth.models.User through SQLAlchemy.

Import and initialize the sabridge.Bridge:

>>> from sabridge import Bridge
>>> bridge = Bridge()

We use the model’s class to obtain the SQLAlchemy version of the table:

>>> from django.contrib.auth.models import User
>>> table = bridge[model]

The sabridge.Bridge returns an instance of sqlalchemy.schema.Table. If we write data in Django, we can then view that data via SQLAlchemy:

>>> User.objects.create(username='alice')
>>> result = list(
>>> len(result)
>>> result[0][table.c.username]



sabridge does not re-use Django’s connection to the database, thus if executing in a transaction, any data modified by either Django or SQLAlchemy will not be visible to the other, until the transaction is committed.

Practically, this means that any test cases that uses both Django and SQLAlchemy will have to inherit from django.test.TransactionTestCase instead of the more typical django.test.TestCase. The TransactionTestCase does not wrap each test in a transaction, thus the data modified by SQLAlchemy and Django is not isolated. Unfortunately, the TransactionTestCase is significantly slower than the normal TestCase. Refer to the TransactionTestCase documentation.


sabridge uses SQLAlchemy’s reflection (autoload=True) to discover the schema of the requested Django model.