A data engineering team aims to maintain high data quality in Databricks Delta tables by ensuring only valid data is written, rejecting any that fails to meet validation criteria. Which Databricks feature or option can best help achieve this goal?