소개
This skill empowers data engineers to maintain high data integrity by providing a structured framework for comparing dbt models. It offers reusable SQL patterns for row count verification, primary key uniqueness checks, and business-critical metric comparisons (like ARR or revenue). Designed for production workflows involving Snowflake and dbt, the skill includes specialized logic for handling billion-row tables and interpreting acceptable variances caused by timing or rounding. It ensures that 'verified' models remain functionally equivalent to 'scratch' models, reducing the risk of logic errors during complex refactoring tasks.
