[Paraview] Loading sequential VTK files all at once

[SamuelKey]

Kent--

Thanks for reading and commenting on my "advice."  I now use EnSight 
binary output to be read by ParaView -- lots of pluses for me.

With regard to *.pvd files and the part="nnn" token, the name="abc" 
token is as yet undocumented, but David King did a nice job of hooking 
it up. To wit, the face sheet for Extract Parts (in version 2.4.4 now 
deprecated?) used to use the part numbers from the part="nnn" token, at 
least I thought so. When one has more than two parts (possible 10's of 
parts), staring at a list on numbers from 1 to 38 on the face sheet and 
trying to decide which ones you wanted to extract for "preferential" 
display processing required a crib sheet. The name="abc" token gave user 
names to each part in place of part numbers -- a much better solution.

I see now for a *.pvd file --

1) Version 2.4.4: Extract Parts filter: name="abc" replaces n's; nice touch.

2) Version 2.6.1: Extract Parts filter: extinct, no more, ...

3) Version 3.2.0: Extract Datasets filter: regresses to integers, that 
is, ignores part="n" and name="abc", but appears to add sub-blocks to 
each Group -- no clues on how to specify multiple blocks constituting a 
Group however.

4) Version 3.2.0: Extract Group filter: lists itself as ExtractLevel and 
ignores part="n" and name="abc", only operates on a range n-m.

I am not complaining, I love and count on ParaView.  Large, 
multi-material meshes are my world and the ability to look at a sub-set 
of objects for which I once used the VTK's XML-formats with parts and 
now use the EnSight part-construct are one of my favorite ParaView 
sub-setting abilities.

We are talking about final polishing issues, and it is too soon for 
Version 3.n.m to be subjected to these kinds of requests.

 From watching the mailing lists, for some reason individuals new to 
ParaView seem to want to start with the early VTK file-formats for their 
data (if not csv-formats?).  A good connection between the possible data 
tokens in a *.pvd file and the filter face sheets could, in my view, 
provide a good experience for new users. They can then move to the 
"stronger" ExodusII and EnSight readers/formats (and possibly other 
strong readers of which I am not aware).

Once again, thanks for commenting. I don't mind going back and 
reviewing/refreshing my knowledge base.

--Sam


Only being a skilled in Fortran-90/95 coder, studying filters in 
ParaView does not for me yield a lot of insight.

Kent Eschenberg wrote:
> SamuelKey wrote:
>> 1) At any given time step, part="nnn" needs to count/id the components 
>> for the model. Each part (for example, Trace_Path_1.vtk) must be a 
>> complete, displayable graphics object.
> 
> FYI: When making some changes to the PVD reader in 2.6.1 I became quite 
> familiar with this part of the code. I never did find anything that used 
> the "part" attribute. Of course, its a good idea to set this attribute 
> anyway so that it is easier to read the pvd file. And this attribute 
> might be used in the future.
> 
> Kent
> Pittsburgh Supercomputing Center
>