03. Using vars & config
️ What are vars and config in DBT?
In DBT, vars (variables) and config (configuration) are used to:
- Make models dynamic
- Reduce hardcoding
- Enable environment-specific customization
Using vars in DBT
Define Variables in dbt_project.yml:
vars:
default_schema: analytics
execution_date: '2024-09-01'
Use in SQL Models:
SELECT *
FROM {{ ref('orders') }}
WHERE schema_name = '{{ var("default_schema") }}'
Default Values:
{{ var("unknown_var", "default_value") }}
Passing Vars at Runtime
You can override variables via CLI:
dbt run --vars '{"execution_date": "2024-12-31"}'
Using config() in Models
The config() function sets model-level configurations inside SQL files:
{{ config(
materialized = 'incremental',
unique_key = 'id',
tags = ['finance', 'incremental']
) }}
This replaces YAML settings in dbt_project.yml for that model.
Common config Options:
materialized → view, table, incremental, ephemeral
unique_key → For incremental models
tags → For grouping models
persist_docs → For documentation
Example: Dynamic Config with Macros
{% macro dynamic_config(materialization) %}
{{ config(materialized=materialization) }}
{% endmacro %}
{{ dynamic_config('table') }}
Best Practices
Use vars for environment-specific settings
Use config() for model-level customization
Combine vars + macros for flexibility
Document all variables in dbt_project.yml
Debugging Tips
Check applied configs:
dbt compile
List all variables:
dbt debug --config-dir