johann sorel ha scritto:
> Here is a test case :
>
> CREATE TABLE places
> (
> ref_id character(8) NOT NULL,
> emplacement geometry,
> CONSTRAINT "mainKey" PRIMARY KEY (ref_id)
> )
> WITH (OIDS=FALSE);
> ALTER TABLE places OWNER TO admin;
>
> //dont forget to add the column in the geometry_columns
>
> insert into places values('SCR08123', GeometryFromText('POINT(1 3)',4326));
>
> CREATE OR REPLACE VIEW myview AS
> SELECT places.ref_id, places.emplacement
> FROM places;
>
> ALTER TABLE myview OWNER TO admin;
>
>
> final Map params = new HashMap<String, Object>();
> params.put("dbtype", "postgis");
> params.put(PostgisDataStoreFactory.HOST.key, "...");
> params.put(PostgisDataStoreFactory.PORT.key, 5432);
> params.put(PostgisDataStoreFactory.SCHEMA.key, "public");
> params.put(PostgisDataStoreFactory.DATABASE.key, "...");
> params.put(PostgisDataStoreFactory.USER.key, "...");
> params.put(PostgisDataStoreFactory.PASSWD.key, "...");
> DataStore store = DataStoreFinder.getDataStore(params);
> FeatureSource fs = store.getFeatureSource("myview");
> fs.getBounds(); //RAISE ERROR
>
> First try to connect to the Table, it works
> Next try to connect to the view and you will have a :
> Exception in thread "AWT-EventQueue-0" java.lang.NullPointerException
> at com.vividsolutions.jts.geom.Envelope.init(Envelope.java:254)
> at com.vividsolutions.jts.geom.Envelope.<init>(Envelope.java:186)
> at
> org.geotools.geometry.jts.ReferencedEnvelope.<init>(ReferencedEnvelope.java:203)
>
Hum, I run the above code and fs.getBounds() does not raise an
error, it return null instead. Are you trying to build a referenced
envelope out of it in your code?
Mind that if you do, you have to deal with null values,
FeatureSource.getBounds() javadoc says that the method can return null
if there is no optimized way to compute the bounds.
What we do in GeoServer is that when that happens, we fall back
on an explicity computation by scanning manually all the features
(which might be very expensive, but in our use case, acceptable
since it's a one time operation, then the bounds get cached).
Also mind that computing the bounds can literally take minutes
even using the optimized methods, I have a table with 20 million
roads where the optimized getBounds() runs for 2 minutes straight
before returning a result.
Anyways, you're right that getBounds() fails for a view, there is
a design problem in the postgis data store. In practice, postgis
ds does not have a stand alone FeatureSource, only a FeatureStore
has been coded. When the primary key is missing (as in a view)
a plain JDBCFeatureSource is returned, that knows nothing of
postgis specifics, and that one always return null to getBounds().
If you add the oid column the datastore will think you're using
a old postgres with oid support, assume that one can be treated
as a primary key, and return an optimized PostgisFeatureStore
instead.
Fixing this would require major surgery, that time is better
spent porting the postgis data store to the new jdbc architecture
instead. I've created a jira issue so that this kind of
test will be added to the jdbc-ng standard test suite:
http://jira.codehaus.org/browse/GEOT-2067
Cheers
Andrea
--
Andrea Aime
OpenGeo - http://opengeo.org
Expert service straight from the developers.
-------------------------------------------------------------------------
This SF.Net email is sponsored by the Moblin Your Move Developer's challenge
Build the coolest Linux based applications with Moblin SDK & win great prizes
Grand prize is a trip for two to an Open Source event anywhere in the world
http://moblin-contest.org/redirect.php?banner_id=100&url=/
_______________________________________________
Geotools-devel mailing list
Geotools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel
