You would use Snowflake over Databricks when your primary need is a robust data warehousing solution with easy SQL access and advanced features like time travel. A specific scenario could be when a business needs to perform complex analytical queries on large datasets, regular reporting, and requires the ability to easily revert to previous states of the data without extensive setup or management, all while having different teams accessing the data simultaneously with separated workloads.
You would use Databricks when you need a unified analytics platform that combines data engineering and machine learning. For example, if you are building a data pipeline that requires streaming data analytics while also developing machine learning models in a collaborative environment, Databricks would be the better choice.