What is dbt (data build tool)

Definition

dbt (data build tool) is a transformation framework that lets analysts and engineers define data models as version-controlled SQL select statements, which dbt then compiles and runs inside the warehouse, adding testing, documentation, and dependency management to the transform layer.

« Back to Glossary Index

Brings software engineering practices, version control, testing, and review, to SQL transforms
Builds a dependency graph of models so the warehouse runs them in the correct order
Generates data tests and documentation directly from the transformation code
Lets analysts own transformations in SQL without managing pipeline infrastructure

Real World Example

An analytics team uses dbt to define layered models from raw tables to a clean revenue mart, with automated tests that fail the build if any order is missing a customer, keeping the warehouse logic tested and documented.

FAQs

What does dbt actually do?

It compiles SQL models and runs them inside your warehouse in dependency order, handling the transform step of ELT with testing and docs.

Does dbt move or store data?

No, dbt only transforms data already in the warehouse; it relies on the warehouse for storage and compute.

Why is dbt popular for analytics engineering?

It applies version control, testing, modularity, and documentation to SQL, making transformations reliable and maintainable.