source() function for raw tables in dbt

Introduction

The source() function helps you tell dbt where your raw data tables are. It connects your project to the original data so you can use it safely and clearly.

When you want to use raw data tables from your database in your dbt models.

When you want to document and track the source of your data for clarity and auditing.

When you want to build transformations on top of raw tables without hardcoding table names.

When you want to make your dbt project easier to maintain by referencing sources consistently.

Syntax

dbt

source('source_name', 'table_name')

source_name is the name you gave to your data source in your sources.yml file.

table_name is the raw table name inside that source.

Examples

This selects all data from the customers table inside the raw_data source.

dbt

select * from {{ source('raw_data', 'customers') }}

This gets the id and name columns from the orders table in the sales_db source, filtering for completed orders.

dbt

select id, name from {{ source('sales_db', 'orders') }} where status = 'completed'

Sample Program

This example shows how to define a source called raw_data with a table customers in the sources.yml file. Then, in a dbt model, it selects the id and email of active customers from that source.

dbt

/* In your sources.yml file */
version: 2

sources:
  - name: raw_data
    schema: raw_data
    tables:
      - name: customers

/* In your dbt model SQL file */
select id, email from {{ source('raw_data', 'customers') }} where active = true

Important Notes

Always define your sources in sources.yml before using source() in your models.

Using source() helps dbt track data lineage and improves documentation.

Summary

source() connects your dbt models to raw tables safely.

It uses names from your sources.yml file for clarity and maintenance.

It helps you write cleaner, more understandable SQL in dbt projects.