[postgis-users] Routing

Daniel Kastl orkney at gmx.de
Mon Jan 28 17:01:52 PST 2008


Hi Andrea,

Maybe the pgRouting Forum is a better place to ask this question:
http://pgrouting.postlbs.org/discussion (For sure I would have answered 
you there, too)
> Kanagawa table provided as example app, but I was guessing what makes 
> a table suitable for being "routed".
For routing with Dijkstra algorithm you only need a "valid" road network 
with information about source (start) and target (end) of each link. 
That's all, and you even don't need something like a geometry column.
>
> Kanagawa has a gid, a source and a target fields, a length field. 
> source and target represents the nodes, while gid represents the 
> branches (obviously there's a geometry column).
>
> Btw, in the tables I found of my town, Padova, I have found no one of 
> these fields or similar, even if who provided it said it was 
> "topologically correct". At this point I am missing what exactly this 
> means, if not that the rapresented geometry is a network of branches 
> and nodes.
If you already have data that has source/target information in the road 
link table, then you already have a valid "network topology" and don't 
need to create it. To be able to use heuristic Algorithm like Astar for 
example, you need geometry information as well.

Did this explanation help?
Daniel

>
> Thanks
>
>
> Andrea Maschio
> http://www.superandrew.it <http://www.superandrew.it/>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> postgis-users mailing list
> postgis-users at postgis.refractions.net
> http://postgis.refractions.net/mailman/listinfo/postgis-users
>   

-- 
Daniel Kastl  Orkney, Inc.
MM Park Building 13F
3-6-3 Minatomirai, Nishi, Yokohama, Japan
TEL +81-45-228-3320  FAX +81-45-228-3321
kastl at orkney.co.jp  www.orkney.co.jp




More information about the postgis-users mailing list