[MayaVi-users] Re: [vtkusers] Re: Strange VTK Behavior...

[Prabhu Ramachandran]

>>>>> "Andrea" == Andrea Gavana <andrea.gavana at gmail.com> writes:

    Andrea> Hello NG, OK, after a day of trying different things, I
    Andrea> have a VTK data file in which only the first 2 vertical
    Andrea> "layers" of my unstructured grid are shown.  The new grid
[...]
    Andrea> All of them shows this problematic behavior... the cells
    Andrea> seems corrupted with severe artifacts, even if I have
    Andrea> checked that there are no coincident cells (at this point
    Andrea> I hope that my checking script is wrong...). The VTK data
    Andrea> file, for those interested in, is here:

I took a quick look and it looks like your data points are very large
in magnitude.  Your surfaces are also very thin.  You are using floats
for the points and data.  That gives you about 8 significant places
and any subtractions involving such large magnitudes will introduce
large loss of significant places.  This is *probably* one issue.  So,
you *might* be better off using "double" instead of "float".  The
other suggestion I have is that you take a look at the surface in
wireframe mode and see if the cells correctly shown.  If not, use the
picker ('p' in mayavi) to determine the faulty point/cell and see if
you can figure out if there is something goofy in your data.  I'm
afraid thats the best I can do with my current time constraints.

cheers,
prabhu