Hi,
I used these sort of tools mostly to check if the data are good (like that every A that should have B indeed have B) and correct them if needed.
I began with WDQ when it was pretty much the only avalaible tool. Right now, I use only SPARQL which is more difficult to master but hundred times more rewarding. NB: I probably couldn't use SPARQL now if I hadn't started with WDQ first, it was a pedagogy tool for me.
For the [dis]advantages, I think a quote by Paul Valéry sum it up : "Everything simple is false. Everything which is complex is unusable." So we must find a middle ground and in the end, I prefer near-unusable to use than near-false.
Cdlt, ~nicolas