Anyone with experience has some detailed info about the limitation, trade-offs and problems of using Datalog instead of SQL?
Datalog seems like a natural choice as declarative query language but has not yet become mainstream. I wonder why?
That's not to say that using the SQL API is a bad idea, but if you're using Crux from Clojure, you'd be missing out on some stuff.
One can make one solve problems of the other but they're really different mental models, which impacts the representation of the data needed to enable their use.
For Datalog one needs relationships- a graph- not what SQL calls relations, which is just data. Datalog requires a richer set of opinions- at least conceptually- about the data.
Languages like Datalog will not become mainstream until graph modeling is mainstream.