The {{ ref() }} function is a fundamental Jinja macro in dbt used to reference other models. Its primary purpose is to enable modular data modeling by allowing you to select from the output of another model. When dbt parses a project, it uses these ref calls to automatically infer dependencies between models. This process builds the Directed Acyclic Graph (DAG), which dictates the correct execution order for dbt run. Therefore, ref is utilized both to reference a model within another (its direct action) and to build the dependency graph (its critical outcome).

A. The ref function resolves to a table or view name; it does not execute the SQL query. The surrounding SQL statement (e.g., SELECT) performs the execution.

B. Model configurations, such as materialization or tags, are defined using the {{ config() }} macro or in .yml files, not with ref.

1. dbt Labs Official Documentation

ref function: "The ref function is the most important function in dbt. It's the way you refer to other models... More importantly

dbt uses these refs to automatically build the dependency graph between models. This is what allows dbt to run models in the correct order..."

Source: dbt Labs Documentation

Jinja Functions

ref.

2. dbt Labs Official Documentation

About dbt models: "The ref function is how you reference one model within another. This is a core concept in dbt

as it allows dbt to automatically build a dependency graph of all of your models."

Source: dbt Labs Documentation

Core concepts

dbt models

About dbt models.

3. dbt Fundamentals Course

Building your first dbt models: "Instead of writing the schema and table name

we can use the ref function... When you use ref

dbt knows that your current model depends on the model being referenced. This is how dbt is able to build your DAG and determine the correct order to run your models in."

Source: dbt Labs

dbt Fundamentals Course

Section: Building your first dbt models.