I think focusing on one and optimizing for efficiency is the way to go. They all have their advantages and disadvantages, and making the best possible schema and access path within a single implementation of DB is the way to go. DBs are so much more than people let them to be. You can save so much on dealing with data and consistency by just using your preferred DB to the full extent πΆπΎπ«‘
Discussion
Yes, I'm focusing on pg now. I'm planning to abstract the repository layer out. Interested developers can develop and optimize their favorite database implementation.