How do you handle null values and data anomalies in SQL when building production-grade ETL pipelines at Prudential?
