Upgraded to PostGIS 1.5.3 and get the same problem.
select gid, st_isvalidreason(the_geom) from parcels where not
isvalid(the_geom);
...<lots of non-fatal notice massages>
NOTICE: Ring Self-intersection at or near point -84.5006 38.044
NOTICE: Ring Self-intersection at or near point -84.5064 37.988
NOTICE: Ring Self-intersection at or near point -84.5097 38.0421
NOTICE: Self-intersection at or near point -84.4921 38.0534
NOTICE: Ring Self-intersection at or near point -84.5265 38.0425
NOTICE: IllegalArgumentException: Invalid number of points in
LinearRing found 3 - must be 0 or >= 4
NOTICE: IllegalArgumentException: Invalid number of points in
LinearRing found 3 - must be 0 or >= 4
ERROR: POSTGIS2GEOS conversion failed
********** Error **********
ERROR: POSTGIS2GEOS conversion failed
SQL state: XX000
So it looks like there are at least to issues:
1. open rings
2. invalid number of points in ring
that are causing problems.
Thanks,
-Steve
On 12/8/2011 1:25 PM, Paul Ramsey wrote:
First confirm if PostGIS 1.5.3 is any better.
If it's not, this should not be a major problem to fix.
P.
On Thu, Dec 8, 2011 at 10:15 AM, Stephen Woodbridge
<[email protected]> wrote:
This was loaded via:
ogr2ogr -t_srs EPSG:4326 out.shp in.shp
This complained about something but seems to have passed on at least some of
the problems.
shp2pgsql ...
No complaints there - so maybe this command should be extended to validate
geometry and have a rule for how to handle bad ones like NULL the geom or
try to fix it, etc.
I was able to locate the first issue, and after about 30 mins of trying this
and that was able to locate the bad record and dump with astext and manually
copy the start point to an new end point and fix that.
The problem is I have 503 more records that are not valid and want to fix
them but the tools do not allow me to easily sort out the simple invalid
from the more severe ones because these commands are dying on fatal errors
instead of recognizing it is a fatal condition are reporting it
appropriately. :(
So my first need it to sort the simple and problematic cases. Any ideas on
this?
Thanks,
-Steve
On 12/8/2011 12:17 PM, Paul Ramsey wrote:
Somehow you've snuck an unclosed polygon into the system. They don't
usually get in, because the parser disallows them.
This will make dumping a test data tricky at best...
P.
On Thu, Dec 8, 2011 at 8:42 AM, Stephen Woodbridge
<[email protected]> wrote:
Hi all,
I'm getting this error:
select gid from parcels where not isvalid(the_geom);
ERROR: pt_in_ring_2d: V[n] != V[0] (-83.5001 38.244 != -83.5001 38.244)
********** Error **********
ERROR: pt_in_ring_2d: V[n] != V[0] (-83.5001 38.244 != -83.5001 38.244)
SQL state: XX000
select gid, st_isvalidreason(the_geom) from parcels where not
isvalid(the_geom);
NOTICE: IllegalArgumentException: Invalid number of points in LinearRing
found 3 - must be 0 or>= 4
NOTICE: IllegalArgumentException: Invalid number of points in LinearRing
found 3 - must be 0 or>= 4
ERROR: POSTGIS2GEOS conversion failed
********** Error **********
ERROR: POSTGIS2GEOS conversion failed
SQL state: XX000
"POSTGIS="1.5.1" GEOS="3.2.0-CAPI-1.6.0" PROJ="Rel. 4.6.1, 21 August
2008"
LIBXML="2.6.32" USE_STATS"
This just seems very bad in that isvalid() and st_isvalidreason() should
never fail, IMO.
-Steve
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
_______________________________________________
postgis-users mailing list
[email protected]
http://postgis.refractions.net/mailman/listinfo/postgis-users
