Configuring data sources for dbt

sources.yml describes how to reach tables created by ELT service

   - name: metrika
     database: "{{ env_var('DBT_MSSQL_DATABASE') }}"
     schema: "{{ env_var('DBT_MSSQL_SCHEMA') }}"
     description: Yandex.Metrika
     tags: ['metrika']
     tables:  
       - name: sessions_facts
         identifier: metrika_sessions_facts
       - name: goals_facts
         identifier: metrika_goals_facts
       - name: purchases_facts
         identifier: metrika_purchases_facts
       - name: devices
         identifier: metrika_devices
       - name: goals
         identifier: metrika_goals
       - name: purchases
         identifier: metrika_purchases
  

Mind the database and schema keys. They might vary between projects and deployments so they are configured as environment variables. I will show you where they get actual values a little bit later.

Identifier key is the full name to reference a source table in a database, while name key works like an alias for referencing a table in dbt code. It comes pretty handy if you have long table names.

By labeling sources with tags you can select, run or test certain parts of your DWH, for example you might want to rebuild models depending on a particular source after fixing a bug.

Listing particular sources with dbt ls command looks like this (click to expand):