[ncl-talk] Triangulation selection

[David Brown]

Hi Mathieu,
Sorry for the delay in responding. I am going to take a look at your
examples and see whether there is a fixable issue, or whether further
development is needed to handle your cases. I suspect the latter, but
I am not sure. In any case, we do want to make this work in a
satisfactory manner.
 -dave

On Thu, Oct 29, 2015 at 11:45 AM, Mathieu Dutour
<mathieu.dutour at gmail.com> wrote:
> Dear all,
>
> I have a problem when plotting data from an unstructured mesh.
>
> The problem that happen is that some triangles are omitted on the
> boundary is therefore not as nice as it could be. This happens in
> at least two contexts (files in attachment):
>
> Case 1:
> I select with   mpLeftCornerLatF  and similar a box
> to make a plot. The nodes that are selected are the ones
> inside the box and the triangles used are the ones of those
> nodes. The problem is that some of the surface on the border
> is not filled because the nodes that cover it are outside of the
> box.
>
> Case 2:
> I select with a box a box that contains all the points of the model.
> The problem is that apparently some triangles on the boundary
> are not used and I do not see why. If all triangles were used then
> I would get a collored rectangle. This problem is independent
> of the above as I made sure that all points are used.
>
> In both cases, if I outcomment
>   res2 at sfElementNodes      = f->ele
> the result is somewhat nicer but it does not really solve the essence
> of the problem, which is missing triangles.
>
> Any help on addressing this problem would be much appreciated.
>
>   Mathieu
>
>
>
>
>
>
> _______________________________________________
> ncl-talk mailing list
> ncl-talk at ucar.edu
> List instructions, subscriber options, unsubscribe:
> http://mailman.ucar.edu/mailman/listinfo/ncl-talk
>