[Paraview] periodic data in ParaView

[Jean Favre]

On 09, Jul 2009 03:45 PM, Berk Geveci <berk.geveci at kitware.com> wrote:

>I am afraid there is no direct way. What you can do however, at the
>cost of more memory, is to use the tranform filter to create a copy
>(or more) of your dataset that is shifted and then group the result
>with the Group filter.
 
I have a similar challenge. A single 3D grid which I warp into a
spherical grid. The Z=0 and Z=max planes meet at the meridian. All is
good, iso-contour surfaces for example appear "full". The challenge
arises when doing streamlines. I have two cases:

A) surface-restricted streamlines on constant-radius surfaces.
Streamlines don't cross over the meridian line. I fixed this by merging
the points along the meridian and the surface then appears fully
connected, i.e. there are no "external" triangle edges.

B) 3D-streamlines, and the the streamlines don't cross over the meridian
plane.

Berk's comment just make me think that I could make a second copy of the
spherical grid (a full shallow copy), group the datasets, and use the
multi-block version of the streamer. (I do streamlines on large
multi-blocks datasets where the inter-block connectivity is never given
by the reader, and it works very well. Particles leaving a block find
their new "host" block and continue along the streamlines). So shouldn't
that work with two copies of a spherical grid?

Feature request: I am not aware that vtkMultiBlockDataSet allows the
definition of block connectivity, yet, I have it for other simulation
data, but my reader doesn't know what to do with it. Wouldn't a
multi-block streamer use it if available?

Jean-- 

Swiss National Supercomputing Centre