Honestly, I really like having all my SQL in front of me and tied to a Git repository. This way, I can actually see who made the changes. My space bar has stopped working on this computer, and I’m half asleep on the couch. So I’ll simply say that dbt’s testing functionality is what saves me. Since dbt can run independently in its own warehouse, separate from the business teams’ Looker dashboards, large-scale processing doesn’t impede their workflow. Additionally, we no longer have to manage different servers for transformation processes, which is good. Review collected by and hosted on G2.com.

In February 2022, a single corrupted webhook caused null values throughout our database, and hours later a generic “error” message in the dbt error log pointed to a problem in some obscure location. Also, if a developer uses a poorly constructed join and fails to monitor the warehouse size, our computing costs rise rapidly. My minor complaint is having to view massive YAML files for schema tests, where one misplaced space causes the entire test to fail. On top of that, dbt releases new versions frequently enough that it seems as though the documentation is always lagging behind. Review collected by and hosted on G2.com.