On Wed, Oct 10, 2012 at 12:16 PM, Sandro Santilli <s...@keybit.net> wrote: > > One benefit is making the spatial relationships explicit, so that > you could encode your counties as being composed by the districts, > and the districts would be composed by faces, which would be defined > by (shared) edges. > At that point querying for "all districts in a county" would just be > a matter of listing the components of a "county" TopoGeometry. > > But I guess you're still in the data preparation phase as you still > don't know which district record belong to which county, right ? > In that case you'll need ST_Covers (assuming the input data is > _really_ correct and doesn't contain cases in which the shape > of a district is slightly outside the shape of its county. > > Note that storing the data in a topology would be an excellent way to > have better control on those kind of data problems (often not easily > detectable as invisible to the eye). > > --strk; >
+1 for this approach I have been using it for managing European administrative borders with excellent results! regards p -- Paolo Corti Geospatial software developer web: http://www.paolocorti.net twitter: @capooti skype: capooti _______________________________________________ postgis-users mailing list postgis-users@postgis.refractions.net http://postgis.refractions.net/mailman/listinfo/postgis-users