I had a similar situation in one of the projects. A separate test database is created before starting the tests and the structure of the prod database is copied into it and filled with fake data. Then it is removed. Tests use a mock object to work with a database, where a test database is substituted instead of a prod. Sometimes you need to run non-destructive tests on existing records in a prod or stage database; or if it is not convenient to update the test data - you can switch the used database.

@elysha Yes, this happens Suffering for a long time, ask a question, and then understand how to do it without an answer

I usually make the expected and actual tuples, and I compare them. expected = ('john doe', '1111', 'something', True) actual = ( response ["user"], response ["objects"] [0] ["id"], response ["objects"] [0] ["event"] [0] ["type"], response ["reached"] ) assert expected == actual Run pytest with the -v switch, then it will clearly show which elements in the tuples are not equal.