Use dbt macros to encapsulate reusable SQL logic and call them across multiple models

Steps

1. Create a macro by adding a .sql file in the macros/ directory of your dbt project; define the macro using Jinja syntax: {% macro macro_name(arg1, arg2='default') %} ... SQL logic using {{ arg1 }} ... {% endmacro %}
2. Call the macro from a model or another macro using {{ macro_name(arg1_value, arg2='override') }}; dbt compiles the macro call inline at the call site — the result is SQL, not a function call at runtime
3. Use macros for logic that varies by adapter (e.g., date truncation, string aggregation): check the target adapter with {{ target.type }} inside the macro to emit database-specific SQL while keeping model code clean
4. Use the dbt_utils package for common pre-built macros (e.g., dbt_utils.date_trunc, dbt_utils.get_column_values, dbt_utils.pivot); install via packages.yml and dbt deps before referencing
5. Test macros by creating a model that calls the macro and running dbt compile to inspect the generated SQL in target/compiled/<project>/<model>.sql before running it against the warehouse