[Paraview] XDMF reads as Multi-group Dataset

[Jerry Clarke]

Eric,

Eventhough I'm a Terp fan .....

The DataItem under Topology allows for XDMF to use part
of the total Geometry. For example if your Geometry specified
100 points your PolyVertex Topology might only use 10 of them.

What "should" happen is that if there is no "DataItem" under
Topology, XDMF generates a "default" one. So if you have 5 points
it would generate "0 1 2 3 4". This seems to be busted at
the moment. I'm looking into it now. I'll get back to you
once I figure out what is happening.


Jerry Clarke




Eric Monson wrote:
> Hey Dave,
> 
> Thanks for the reply.
> 
>  > Berk can correct me if I am wrong, but that is a limitation of
>  > Paraview. Once an atomic dataset is put into a composite data set, it
>  > can't be pulled out back into the atomic dataset's type. We hope to
>  > have it fixed for the upcoming 3.4 release. The new XDMF reader, which
>  > I hope to have ready in a week or two, will produce atomic datasets
>  > when there is only a single grid in the file.
> 
> It sounds like the issue you are trying to address with the new XDMF 
> reader is at the root of my problem -- XDMF files containing a single 
> grid are being read in to ParaView as a Multi-group dataset. Right now 
> I can't get any XDMF file to read in as a pure type.
> 
>  > This sounds like a bug in the new XDMF library. You might want to ping
>  > the xdmf mailing list (xdmf at lists.kitware.com) to see if any one is
>  > aware of the issue, has a workaround or is working on a solution.
>  > Perhaps the ARL group knows more than I.
> 
> Thanks for letting me know about the XDMF mailing list. I hadn't seen 
> any reference to that before.
> 
> Reading through the XDMF v. 2.0 Model & Format page again, I noticed 
> the sentence: "For unstructured topologies the Topology element must 
> contain a DataItem that defines the connectivity." This doesn't make 
> sense (to me :) for Polyvertex topologies, which they define as "a 
> group of unconnected points". I'll see if anyone can give me more 
> insight on that, or suggest a workaround.
> 
> Talk to you later,
> -Eric
> 
> _______________________________________________
> ParaView mailing list
> ParaView at paraview.org
> http://www.paraview.org/mailman/listinfo/paraview
>