[postgis-users] Shared geometries?

[Dylan Beaudette]

what about doing the topological stuff in GRASS?

cheers,

dylan

On Friday 11 May 2007 16:19, Stephen Woodbridge wrote:
> Brent,
>
> This is probably the best strategy that might work given the current
> state of postGIS. I think the big issues for most people is that there
> currently is not a lot (any?) support for topology in postGIS today that
> would hide some or all of the nuts and bolts of building a topological
> net for a large number of points, lines and polygons.
>
> I tend to think of these tasks in terms of manipulating all of the
> Census Tiger data. There are 3300 counties in the US so it is quite a
> large task to set this model up for 50 states and 3300 counties. Having
> some tools to automate building the topology, checking that it is
> consistent and isValid on changes are important to include in a model
> that supports topology. I would love to see progress in this area,
> especially, because US Census will soon be releasing Tiger data in
> shapefiles which will make it easier to load and work with in postGIS.
>
> but I digress ...
>    -Steve
>
> Brent Wood wrote:
> > As mentioned in other replies, this requires a topological model which,
> > unless I'm missing something, is feasible now with PostGIS, as described
> > below.
> >
> >> Suppose I have a shapefile containing county data, with polygon
> >> geometries representing the county boundaries.  And another shapefile
> >> containing state data, with polygon geometries representing the state
> >> boundaries.
> >>
> >> The state boundaries and the county boundaries should coincide, though I
> >> know of no way of enforcing that with shapefiles.
> >>
> >> Is there a way, in PostGIS, to create two layers that have different
> >> sets of attributes, but whose gemoetries share the same points?  So that
> >> if I edit the county boundaries, the state boundaries will automatically
> >> reflect those changes?
> >
> > If you can generate the constituent linework required to define the
> > polygons, you could (theoretically anyway) have a table of unique lines,
> > then tables identifying the lines for each county & state, & build the
> > polygons from the lines as required. This would pretty much be an
> > implementation of a topological model. Edit the underlying lines & all
> > dependent polygons are automatically updated.
> >
> > Perhaps instantiate the polygons as views on the lines table....
> >
> >
> > t_lines ---< t_county_line >--- t_county
> >
> > line_id      line_id            county_id
> > line         county_id          name
> > ...                             ...
> >
> >         ---< t_state_line >--- t_state
> >
> >              line_id            state_id
> >              state_id           name
> >                                 ...
> >
> >
> > Then a view polygonising the lines for each county, or if you prefer,
> > maybe a cron job or trigger to build a table with the polygons every day
> > or on a change to the lines.
> >
> > I guess the hardest part would be generating the initial linestrings.
> >
> >
> > Cheers
> >
> >   Brent Wood
> >
> > _______________________________________________
> > postgis-users mailing list
> > postgis-users at postgis.refractions.net
> > http://postgis.refractions.net/mailman/listinfo/postgis-users
>
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users

-- 
Dylan Beaudette
Soils and Biogeochemistry Graduate Group
University of California at Davis
530.754.7341