Hi,
I would like to discuss what seems to be a common bug in our codebase.

By definition FeaturesSource.getBounds(query) returns the bounds of
the collections that would be returned by the query.

Now, the filter is honored, so only the bounds of the features that
are actually catched by the filter is returned, but the properties are 
usually forgotten about.

If the Query q extracts only non geometric attributes, what should be 
the result of fs.getBounds(q)? Imho they should be empty, as if
the feature type was geometryless, because the feature collection
returned by fs.getFeatures(q) _is_ geometryless.
Yet the database code atm just ignores that and returns non empty
bounds using the default geom or all the geoms.

What if we have a feature type with multiple geometries and q extracts
only one of them? Imho the fs.getBounds(q) should return the bounds
of just the geometry column extracted. Yet, the current code either
returns the bounds of the default geometry or uses them all.

What if we have multiple columns that are in different srs?
The resulting bounds are one, imho everything should be reprojected
to the srs of the first geometric column requested. The alternative
would be to use the crs of the default geometry, but that seems
wrong, if you request the bounds of a single geometry column
in a SRS other than the default one, wouldn't you expect the
bounds to be returned in the SRS of the requested one?

Opinions?

Cheers
Andrea

-- 
Andrea Aime
OpenGeo - http://opengeo.org
Expert service straight from the developers.

------------------------------------------------------------------------------
Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day 
trial. Simplify your report design, integration and deployment - and focus on 
what you do best, core application coding. Discover what's new with 
Crystal Reports now.  http://p.sf.net/sfu/bobj-july
_______________________________________________
Geotools-devel mailing list
Geotools-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geotools-devel

Reply via email to