[Qgis-developer] Suppress rendering of small features

[Mathieu Pellerin]

Could this be done using a rule-based expression through the creation of a
variable that returns the pixel-size area of a polygon / pixel-size length
of a line?


On Thu, Sep 12, 2013 at 3:28 PM, Matthias Kuhn <matthias.kuhn at gmx.ch> wrote:

> Hi,
>
> How would you define small? I guess something can already be done with
> scale based visibility, but that needs preprocessing of the data and
> assigning some kind of "size class".
>
> On Don 12 Sep 2013 08:51:43 CEST, Andreas Neumann wrote:
> > Hi,
> >
> > Another interesting thing would be to simplify features when you zoom
> > out. Maybe simplifying would be quicker than rendering thousands of
> > unnecessary vertices. Optionally this could also be "outsourced" to the
> > database - many databases support simplification on the db level. I
> > don't know if it would really speed up things. One would have to test.
>
> I did a preliminary analysis for this.
> I created a view on a postgis DB containing ~17000 line features with
> the geometry column simplified to 10m in the view.
>
> *Non Simplified*
> ./output/bin/qgis_bench --iterations 20 --width 1024 --height 768
> bench-full.qgs
> iterations: 20
> total_avg: 0.4246
> total_max: 0.657
> total_maxdev: 0.2324
> total_min: 0.407
> total_stdev: 0.0535176606364665
>
> *Simplified*
> ./output/bin/qgis_bench --iterations 20 --width 1024 --height 768
> bench-simplified.qgs
> iterations: 20
> total_avg: 0.09615
> total_max: 0.098
> total_maxdev: 0.00184999999999998
> total_min: 0.095
> total_stdev: 0.000792148975887722
>
> 0.42 / 0.09 is more than 400% performance improvement for this simple
> case.
>
> The results of course depend heavily on the particular setting, but I
> think this looks indeed like a road we should have a look into.
>
> (Setup: PostGIS 2.0.3 in a kvm virtual machine on the same physical
> machine. No styling at all took place on the layer)
>
> Matthias
>
> >
> > Andreas
> >
> >
> >
> > Am 12.09.2013 07:57, schrieb aperi2007:
> >> Not always is preferrable to suppress the small features.
> >> Often in our rendering we prefere to substitute a small rendering of
> >> kind polygon with a point rendering at lowest scales.
> >>
> >> This is possible using the Rule-Rendering
> >> also is possible to use the rule-rendering to define a rule that in a
> >> scale interval and with a dimension interval the feature is render using
> >> a point rendering and change in a polygon rendering when the scale grow.
> >>
> >> Andrea.
> >>
> >> On 12/09/2013 02:13, Nyall Dawson wrote:
> >>> Hi all,
> >>>
> >>> I've been chatting with Nathan about adding a feature to QGIS which
> >>> would allow suppression of rendering features smaller than a set
> >>> threshold for vector layers. (There's already a similar function in
> >>> the labelling engine "Suppress labeling of features smaller than").
> >>> This feature would come in handy for me for several tables which
> >>> contain a wide range of very small -> very large features, which cause
> >>> slow rendering at small scales.
> >>>
> >>> I'm just checking that if I code this up it's not going to be
> >>> duplicated effort - Nathan is under the impression that someone else
> >>> has been talking about adding something similar. Is there currently
> >>> any plans for a feature like this, or should I get started?
> >>>
> >>> Nyall
> >>>
> >>>
> >>>
> >>>
> >>>
> >>> _______________________________________________
> >>> Qgis-developer mailing list
> >>> Qgis-developer at lists.osgeo.org
> >>> http://lists.osgeo.org/mailman/listinfo/qgis-developer
> >>
> >>
> >>
> >>
> >> _______________________________________________
> >> Qgis-developer mailing list
> >> Qgis-developer at lists.osgeo.org
> >> http://lists.osgeo.org/mailman/listinfo/qgis-developer
> >>
> >
> > _______________________________________________
> > Qgis-developer mailing list
> > Qgis-developer at lists.osgeo.org
> > http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
>
> _______________________________________________
> Qgis-developer mailing list
> Qgis-developer at lists.osgeo.org
> http://lists.osgeo.org/mailman/listinfo/qgis-developer
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20130912/be94b87b/attachment.html>