I am writing a REST API using flask_restful and managing the mysql db using flask-sqlalchemy. I would like to know what the best practice for loading existing data into a table when the app starts is.
I am currently calling the db.create_all() method withing an endpoint with the @app.before_first_request decorator. I would like to then fill in one of the tables created with existing data from a csv file. Should the code to push the data in a separate script or within the function?
Thanks!
I would separate loading initial database data from application initialization, because probably initial data from my experience would not be changed often and can take some time if file is bigger, and usually you don't need to reload it in the database each time application loads.
I think you will most certainly need database migrations at some point in your application development, so I would suggest setting up Flask-Migrate to handle that, and running its upgrade method on application creation (create_app method if you are using Flask application factories pattern) which will handle database migrations. I am saying this since it will save you some headache when you are introducing it later on database already populated with actual data which is initialized with db.create_all().
And for populating database with seed data I would go with Flask CLI or Flask-Script. In one of my recent projects I used Flask-Script for this, and created separate manage.py file which amongst other application management methods contained initial data seeding method which looked something like this:
@manager.command
def seed():
"Load initial data into database."
db.session.add(...)
db.session.commit()
And it was run on demand by following command:
python manage.py seed