Frequently Asked Questions

How does tuberia compare with dbt?

dbt is more SQL-centric and handles queries as templates. In tuberia tables are classes that can execute arbitrary Python code (including SQL if for example you are using PySpark).

To do unit tests in dbt you must learn how to define them and where to put them (you must define a macro that allows you to change where the table you want to test takes data from, you need to add a separate file with the test data...). In tuberia, if you know how to test in Python you know how to test the classes that create tables (mocks to the rescue!).

In dbt dynamic queries can get complicated with the use of Jinja templates. When it comes to dynamic code there is nothing simpler than using a programming language to define transformations.

Note

Using arbitrary Python code in dbt is coming but it is still to early to compare.

How does tuberia compare with Prefect?

Tuberia is not an orchestrator. Tuberia can be seen more as a compiler, passing Python code that anyone with Python skills can handle, to a Prefect flow (or to any other orchestration tool) that you can deploy in your environment.

How does tuberia compare with Airflow?

Tuberia is not an orchestrator. Tuberia can be seen more as a compiler, passing Python code that anyone with Python skills can handle, to an Airflow DAG (or any other orchestration tool) that you can deploy in your environment.