Hi Frederic,
I think this is the case in other wrapper functions as well, because Shooting Star core returns a list of road links and the wrapper(s) just modify the output to return the geometry column, for example to draw the line.
This geometry is the one of your data and how the data was digitized. And some attributes may be different depending on the direction, best example are one way streets.
So it’s a good question if the wrapper should care about this and flip the linestring, or if this should be up to the application to check for the right order of start and end of each road link. Because there might be direction dependent attributes that we don’t know about and might go beyond a general function.
At the moment you have to care about this. But we can talk about if it should be done by pgRouting in the future. Maybe you can extend the shootingstar_sp_smart function to do it right.
Just as a side note, the first and last edge are added and don’t exist in the network. So they are in the right direction because of that. Since the links don’t really exist, they have no attributes either.
Daniel
2011/6/23 Frederic Junod <frederic.junod@camptocamp.com>
Hello,
I’m using the shootingstar_sp_smart function from
https://github.com/pgRouting/pgrouting-contrib/ and I’m having an
issue with the “direction” of the result.
For example, the result of a query (from x to y) is:
x ---- a ---- b ---- c ---- y
The function return something like:
MULTILINESTRING((x, a))
MULTILINESTRING((b, a))
MULTILINESTRING((c, b))
MULTILINESTRING((c, y))
The first and last line are ok but the two other are reversed.
Is my network ok ?
Regards,
fredj
–
Frédéric Junod
Camptocamp SA
Pgrouting-users mailing list
Pgrouting-users@lists.osgeo.org
http://lists.osgeo.org/mailman/listinfo/pgrouting-users
–
Georepublic UG & Georepublic Japan
eMail: daniel.kastl@georepublic.de
Web: http://georepublic.de