over the next few months @obelisk is refactoring the internal storage representation for table content, once that's complete we can explore pre-calculating CRCs and (1) speeding up the set different for the remaining non-keyed tables, but also providing that as a hidden column for folks to add to scheduled queries