Update: added two detectors that don't need this project's own planted bugs -- detect_missing_ownership and detect_undocumented_upstream fire on any real DataHub catalog. Checked them against DataHub's own real showcase-ecommerce sample data: 47 of 67 real datasets had no owner recorded, 53 of 67 had no description anywhere (checked DataHub's separate documentation aspect too, so this isn't an artifact of the wrong field). The pattern is more interesting than the raw number -- governance holds at the dbt source of truth (12/13 owned, 13/13 documented) and evaporates across all 54 downstream copies in Snowflake, Postgres, Looker, PowerBI, S3, and Tableau (8/54 owned, 1/54 documented). Full breakdown in examples/real_catalog_validation.md.
Log in or sign up for Devpost to join the conversation.