dbt-macros

Cross-database dbt macros use adapter.dispatch patterns to route execution to dialect-specific implementations. This ensures your Jinja macros generate correct SQL across different warehouses by overriding built-in dbt macros for naming and aliasing conventions.

Your dbt macro lacks execute guards. You must wrap run_query and DDL execution in an execute variable check so warehouse queries only run when dbt is executing, preventing unintended database calls during compile or docs generation.

Adapter.dispatch overrides built-in dbt macros by routing Jinja macro calls to dialect-specific implementations. You create default and warehouse-specific macro versions, allowing dbt to select the correct SQL generation logic based on the active adapter.

Writing dynamic SQL macros requires correct handling of dbt context variables including this, target, model, and execute. These variables provide environment-aware behavior, schema and alias overrides, and safe execution state checks for your Jinja logic.

You can use run_query safely by adding execute guards to your dbt macros. This prevents warehouse queries from running during compilation or docs generation, ensuring dynamic SQL execution only occurs when dbt is actively running models.

Common production-grade dbt macro patterns include dynamic pivots, unioning relations by column alignment, audit columns, grants, idempotent DDL, and column-existence checks. These reusable Jinja building blocks reduce fragile copy-paste SQL and hardcoded dialect logic.