data-quality

name	data-quality
description	Add data quality checks and tests to dbt models and pipelines.
metadata	{"openclaw":{"always":true}}

Data Quality

You help add data quality checks to dbt models and data pipelines.

Instructions

1. Analyse the model -- read the SQL and understand the grain and business keys.
2. Recommend tests based on column type:
   • Primary keys: unique, not_null
   • Foreign keys: relationships
   • Enums/status fields: accepted_values
   • Dates: not_null, range checks
   • Amounts: not_null, non-negative checks
3. Generate schema YAML using templates/dbt/schema.yml.j2 with tests.
4. Custom tests: When generic tests are insufficient, write singular tests (SQL files in tests/) or custom generic tests (macros).
5. dbt expectations: Suggest dbt_expectations or dbt_utils tests where appropriate (row count, expression checks, etc.).
6. Freshness: Add source freshness checks to source.yml for critical sources (loaded_at_field, warn_after, error_after).

Common Patterns

• Every staging model should have unique + not_null on the primary key
• Mart models should have referential integrity tests
• Incremental models should test for duplicate records after merge
• Add dbt_utils.expression_is_true for business rule validation