Given a specific dataset, how would you approach an ETL (Extract, Transform, Load) task to ensure data integrity?
