[postgis-users] problem clipping a feature
Andre Oliveira
andreolivei at gmail.com
Wed Sep 2 08:23:11 PDT 2009
Hi Everybody,
I was trying to use the ST_Intersection function to overlay (clip) a poygon
by another, with the following code from the workshop at foss4g2007:
CREATE TABLE pg_voting_areas AS
SELECT
ST_Intersection(v.the_geom, m.the_geom) AS intersection_geom,
ST_Area(v.the_geom) AS va_area,
v.*,
m.name
FROM
bc_voting_areas v,
bc_municipality m
WHERE
ST_Intersects(v.the_geom, m.the_geom) AND
m.name = 'PRINCE GEORGE';
After this operation I restarted pgAdminIII, generated an index for
pg_voting_areas, run vacuum analyze, and defined GID as the primary key for
pg_voting_areas.
But when I try to connect with QuantumGIS to visualize the data, two
datasets where recognized with the same name, one bearing the expected
result, but the other a display with more objects than expected for a clip.
Then I tried to inserted the dataset's line into table geometry_columns:
INSERT INTO geometry_columns (f_table_catalog, f_table_schema,
f_table_name,f_geometry_column, coord_dimension, srid, type)
SELECT '', 'public', 'pg_voting_areas', 'the_geom',
ST_CoordDim(the_geom),
ST_SRID(the_geom),
GeometryType(the_geom)
FROM public.pg_voting_areas LIMIT 1;
After I did this, one of the datasets recognized by QuantumGIS disappeared,
but it was the undesired one that survived. Also, I noticed that some of the
measurements in the workshop are not totally equal to the ones I obtain (for
instance, when I query the clipped poygon's area and the clipping
municipality's area, they are not exactly equal - 326579103.825073 versus
326579103.825928).
I am using the last versions of postgre, postgis and quantumgis.
When I installled postgre I chose locale by default and created the postgis
database manually, bearing UTF8 encoding.
I would really appreciate a hand here.
Thanks
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/postgis-users/attachments/20090902/f8d6eadf/attachment.html>
More information about the postgis-users
mailing list