sumitPanda
Databricks Partner

I would always use the UC way. That is the standard for the Production workloads. 

Sumit Panda