It is very nice that dbt gives us the capability to write out our transformation logic as application code. The process of integrating it with Git implies that our whole team may contribute to data models and track changes without hassles, and revert changes as well. Even the separation of the development and production environment out of the box is a tremendous advantage, so we can ensure that we are testing new things without necessarily influencing live reports. Lastly, the automated documentation generation process has made the data models so simple to comprehend, even to the business users, without ever having the touch a SQL. Review collected by and hosted on G2.com.
The initial set up of dbt can be a little fiddly at times and with a mix of operating systems and local database connections. It also lacks its own scheduler, so still we will require an external orchestration tool to carry out our data pipelines. The more advanced Jinja templating and macros proved to be more cumbersome to our team to learn than the simpler SQL. Review collected by and hosted on G2.com.
