It might be advantageous to use scisql support for fast point-in-region searches when executing point-in-polygon (and similar) searches via Qserv.

Since Qserv already limits the execution of such queries to just those chunks overlapping the search region, it may not be worth doing this (unless chunks are really large).

This very low priority issue corresponds to ticket 2056 from Trac, and serves as a reminder that we should do some performance testing to figure out whether to pursue this potential optimization or not. In particular, testing performance on a 1 million row chunk that is 60-80% covered has been suggested.