Tworzenie schematu za pomocą create_all

Your models describe tables, but the database file is still empty. You need one step to turn declarations into real tables. 🏗️

db.create_all() scans every model it knows and issues the CREATE TABLE statements for you. One call builds the whole schema.

create_all must run inside an application context so it can find your config and the right database connection.

create_all only builds tables for models that are imported. If a model file is never loaded, its table silently goes missing.

A clean way to set up the database is the flask shell, where the app context is already active and ready.

create_all skips tables that already exist, so calling it twice will not crash. It simply creates whatever is missing.

Here is the catch: create_all never changes an existing table. Add a new column and the old table stays exactly as it was.

Need a fresh start in development? db.drop_all() deletes every table and its data, so handle it with real care.

A common dev reset is drop then create. It rebuilds clean tables, but it also erases all data, so never do this in production.

With SQLite, a database file like app.db shows up after create_all. Seeing it confirms your tables were actually built.

Because create_all cannot evolve a schema, real projects switch to migrations as soon as the model starts changing over time.

You added a new column to an existing model. What does create_all do?

You built tables with db.create_all() inside an app context, learned it never alters existing tables, and met drop_all. 🎉