Created by @sbalnojan for the finishslime.com & thdpth.com || Updates: Sept/2023 (3x); Aug/2023

Dbt is just SQL + some templates, right?

Yes, and no. It’s become so much more, that we think you need a complete reading guide to get good at it!

Here are the best articles on dbt we came across.

Click on expand on the bullets to get a short description!

But first, let’s bust few dbt myths:

  1. dbt is easy to learn - SQL and Jinja are easy to learn. Dbt might have been years ago, but now it’s a monster.
  2. dbt should run on dbt cloud - transformation is just one step in a large ingest data - transform data - serve data pipeline that should be orchestrated together. The first question should be how to run the whole thing, not dbt alone. ⇒ look into the “4) Deploying dbt” section below!
  3. dbt makes testing easy - Yes, dbt makes testing something easy - and no it doesn’t make writing the right tests easy. ⇒ look into the “3) testing in dbt” section below!
  4. dbt is SQL and thus lock-in free - it was. But now with hundreds of macros, project specific configurations, extensions, huge dbt-internal DAGs its a huge lock-in machine.
  5. dbt is SQL only - Dbt is a lot more: It’s SQL + Jinja + lots of macros + dbt-specific knowledge.
  6. dbt is performant - Dbt is there to bring analysts more capabilities, at the cost of making queries performant.
  7. dbt makes data models transparent - the dbt project docs are fun, but they don’t serve as a “data catalog” unless you put some serious work into it, just as much as into any other.
  8. dbt is without alternatives - weirdly, that was basically true 1-2 years ago, but since has changed a lot; see 6) Potential Alternatives.

1) Learning the dbt essentials

4) Advanced dbt

5) Deploying dbt

2) dbt implemented

3) Testing in dbt

6) Potential Alternatives