[Qgis-developer] Approximation and geoprocessing

[Paolo Cavallini]

Il 13/01/2014 09:12, Bernhard Ströbl ha scritto:

> it depends...
> Processing works as follows: take input layer(s) - do a process - create output layer
> So if the "process" is a selection the result is an output layer with only the
> selected features included.
> Therefore for "eliminate" I included a simple logical selection based on a field in
> the layer. Thus to use eliminate in a model you first have to create an input layer
> with a field that somehow marks all features to be eliminated (e.g. caluclate
> area/perimeter). Currently I do not know how this can be achieved. :-(
> Any hints welcome.

IMHO the approach should be different: eliminating features a posteriori is always
feasible, but very inefficient, and it does not solve the "mouse tail" problem: a
valid feature is created, but attached to it hangs a close-to-zero-area that results
from different approximation of coordinate values.
This may seem a corner case, but is indeed preventing effective analyses in real cases.
All the best.

-- 
Paolo Cavallini - www.faunalia.eu
Corsi QGIS e PostGIS: http://www.faunalia.eu/training.html