Well, actually my scenario can be well covered with the status quo right now. Now that I know how the design philosophy is, it is easy to accept and provide local helpers for special constructs. However, my original solution idea would require something like that, yes. It would require some dynamic translation layer from Pure Clojure Query Representations to dialect-specific SQL queries. The problem is: What would the pure representation look like? That endeavour is the same as creating a new, feature-enhanced SQL standard. I mean a pragmatic starting point would be to just elevate PostgreSQL to the de-facto-standard and translate from Postgres to other dialects.
@johannes.f.knauf Unfortunately PostgreSQL has a lot of features that other databases don't have so there would be a lot of stuff that couldn't be translated to other database dialects.
(and such a library wouldn't come from me since I don't use PostgreSQL)
That is all true. The question is: Do we have a better candidate for some meta-SQL covering a good amount of the common cross-provider query features?