What is ref in dbt: Explanation and Usage

ref in dbt is a function used to reference other models within your project. It helps dbt understand dependencies between models and ensures they run in the correct order.

How It Works

Think of ref as a way to point to another data model inside your dbt project. When you write SQL in dbt, you often need to use data from other models. Instead of typing the full table name, you use ref('model_name'). This tells dbt to find that model and use its output.

This is like giving directions to a friend: instead of saying the full address, you say "go to the house I told you about before." dbt then figures out the exact location and order to build your models so everything works smoothly.

Example

This example shows how to use ref to select data from another model called customers.

sql

select
  id,
  name,
  email
from {{ ref('customers') }}
where active = true

Output

A table with columns id, name, email filtered to only active customers.

When to Use

Use ref whenever you want to build a model that depends on another model's data. It helps dbt know which models to run first and keeps your project organized.

For example, if you have a sales model and want to create a sales_summary model based on it, use ref('sales') inside sales_summary. This ensures sales runs before sales_summary.

Key Points

• ref creates dependencies between models.
• It helps dbt build models in the right order.
• Using ref makes your project easier to maintain and understand.
• It abstracts away physical table names, so you can rename models without breaking code.

Key Takeaways

ref links models and manages build order automatically.

Always use ref to reference other models instead of hardcoding table names.

ref improves project clarity and reduces errors.

It allows dbt to track dependencies and optimize runs.