Thank you for your help. The example is exactly what I need.
I apologize for speaking French to other users. If I again need help or if one day I can bring my own, I will try to do it in English.
Good job for this function Remi and thank you again.
How can i close this topic and note that my problem is solved?
Cheers,
Guillaume

Le 03/02/2015 13:10, Rémi Cura a écrit :
/*french version at the end*/
Ok I think I found the problem :
your geometry are not simple, that is in postgis langage, there is no selfintersection. Your lines have tone of selfintersection , which is forbiden (because GEOS does topological computation)!

So now of to fixe this : use this function <https://github.com/Remi-C/PPPP_utilities/blob/master/postgis/rc_makesimple.sql> to transform your non-simple (multi)line into multiline where each line inside is simple.
You code becomes :

SELECT objectid, nom, code
, ST_OffsetCurve(dmp.geom, 25, 'quad_segs=4 join=round') as offseted_linestring
FROM web.trs_itineraire , st_Dump(rc_MakeSimple(shape)) as dmp
WHERE code = '06-20B'

It execute successfully (now does it do what you expect, I don't know)
(tested on the data you shared, no translation, no snapping)
Cheers,
Rémi-C

~~~~~~~~ French version ~~~~~~~~

J'ai trouvé le problème je pense.
En fait les géométries ne sont pas simple au sens de PostGIS, c'est à dire "qui ne s'autointersecte pas". Cela est interdit et cause les erreurs de GEOS (qui fait du calcul topologique).

Maintenant la solution pour régler simplement le problème :
utilisercette fonction <https://github.com/Remi-C/PPPP_utilities/blob/master/postgis/rc_makesimple.sql> pour passer de (multi)lignes avec des autointersections à des multilignes dont chaque lignes composante ne s'autointersecte pas.

Du coup,
votre code devient :
SELECT objectid, nom, code
, ST_OffsetCurve(dmp.geom, 25, 'quad_segs=4 join=round') as offseted_linestring
FROM web.trs_itineraire , st_Dump(rc_MakeSimple(shape)) as dmp
WHERE code = '06-20B'

La requete s'execute sans erreurs (après le résultat ne sera pas nécessairement ce que vous attendez, mais c'est une autre histoire). (testé sur les données que vous avez partagées, sans translation ni snapping).

Salutations,
Rémi-C

2015-02-03 9:12 GMT+01:00 Sandro Santilli <s...@keybit.net <mailto:s...@keybit.net>>:

    On Sun, Feb 01, 2015 at 08:46:16PM +0100, Rémi Cura wrote:

    > We are coming closer to a GEOS or PostGIS bug.

    Please use the bug tracker for these, thanks.

    --strk;
    _______________________________________________
    postgis-users mailing list
    postgis-users@lists.osgeo.org <mailto:postgis-users@lists.osgeo.org>
    http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users




_______________________________________________
postgis-users mailing list
postgis-users@lists.osgeo.org
http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users

--
Guillaume ARNAUD
Cellule SIGD
Direction de l'Informatique
Conseil Général de Tarn-et-Garonne


_______________________________________________
postgis-users mailing list
postgis-users@lists.osgeo.org
http://lists.osgeo.org/cgi-bin/mailman/listinfo/postgis-users

Reply via email to