I always have considered the Continuous Delivery (CD) for Database world as one of the last frontiers to cross, it is not easy at all and the DevOps guys are brave witnesses about it, however, I consider that it is an aim which can be reachable with a mix of effort, compromise and obviously investment (time and money).

The previous aspects mentioned above are from my point of view the most valuable, but we also need to be realistic about in which stage of this process our companies are, I mean, it is not realistic or fair to pretend that a new startup consider spending many resources for implementing DB deployment under
CD vision, but if the team only requires to deploy through a few environments it can be feasible to use a simplistic approach based on using PowerShell.

I have to confess; I have had traumatic experiences with variant data type in my beginning as a software developer, in fact, I always have believed that it is an anti-pattern, at least in the OOP world, on the other hand in the Database world it gains much relevance, Why? Because of every column inside a table is supposed to be a specific datatype, otherwise, it could create a great number of conflicts and uncomfortable situations.