I have found it unnecessarily distracting to evaluate reputable databases on incidental features. If an important feature is missing it will be added eventuall

I have found it unnecessarily distracting to evaluate reputable databases on incidental features. If an important feature is missing it will be added eventually, if you need a very specific feature then you already have your answer as to which DB to use. Unless you have a very specific reason not to use PostgreSQL as your SQL database then you should just use it. I simplify the selection to the more general architectural value. Do I need SQL, NoSQL, or graph? What about deep JSON queries, embeddability, specific performance requirements, scalability? Given that, I use PostgreSQL as the default choice for SQL/JSON db in all cases that don't require imbedding. SQLite for imbedding. There is one big disadvantage, the same for all fixed schema index based databases. Scalability. Engineering effort rapidly goes to infinity as we asymptotically approach the throughput event horizon.