[postgis-users] postgis on postgresql 8.2 vs 8.3

Hi all, I have been holding back on upgrading my postgis installations from postgresql 8.2 to 8.3 because: 1) it seemed like there was a lot of new postgresql stuff to figure out in 8.3. What are the big changes I need to be aware of? 2) it seemed like there were initially some bugs and iss

Re: [postgis-users] reparing invalid geometry (polygon defined byoneor two points)

Thanks a lot, I will try and get back with result. On Tue, Mar 24, 2009 at 1:24 PM, Obe, Regina wrote: > Ivan, > > Oops more of a variant.   I see you have invalid polygons as well not > just rings.  Hmm try this instead.  I didn't check my logic, but > hopefully you get the idea if this doesn't

Re: [postgis-users] how to correct shared boundary data problem?

Did you try ST_SnapToGrid() to force your geometries to lie on a grid? http://postgis.refractions.net/documentation/manual-svn/ST_SnapToGrid.html Note: this is close to, but not the same thing as, reducing the precision on geometry coordinates. -- Kevin John Zhang wrote: Hi there, I have two

[postgis-users] how to correct shared boundary data problem?

Hi there, I have two feature classes which are supposed to have shared boundaries. However, the data came from another party with the shared boundary data issue, which boundary lines are slightly moved (no snapping during editing?). I am thinking how I can solve the data problem by using PostGIS

Re: [postgis-users] when setting up postgis libs, "could not find function "topologypreservesimplify"" in "...liblwgeom.so"

Thanks for the response, Paul. Yes, I had remembered to do 'make install'. I did find a workable solution, though, so I wanted to post it for the sake of future users who may have similar problems. Basically, I just went into the 'lwpostgis.sql' file and commented out the lines that called the

RE: [postgis-users] reparing invalid geometry (polygon defined byoneor two points)

Ivan, Oops more of a variant. I see you have invalid polygons as well not just rings. Hmm try this instead. I didn't check my logic, but hopefully you get the idea if this doesn't compile. This is a bit of overkill if you have no multipolys with invalid rings. But I'll leave as an exercise fo

RE: [postgis-users] reparing invalid geometry (polygon defined by oneor two points)

Ivan, I think a variant of Simon's script would work for you except instead of keeping rings with a large enough hole, keep rings that are closed and have enough points http://www.spatialdbadvisor.com/postgis_tips_tricks/92/filtering-rings-i n-polygon-postgis so something like CREATE OR REPLACE

[postgis-users] reparing invalid geometry (polygon defined by one or two points)

Dear PostGIS users, we have quite huge MULTIPOLYGON database table containing some 3 percents of invalid geometries. Most problematic to repair are MULTIPOLYGONs which often contain polygon defined by one or two points. Example is here (second polygon is defined by one point): MULTIPOLYGON(((-2