[Paraview] Problem with saving data along stream trace in parallel

[Andy Bauer]

It may be that in serial it traverses the blocks in a separate order or
that the order the point data arrays are added are different in parallel.
An easy check would be to write out the multiblock in parallel before the
streamline filter is used and look at the outputted dataset files. I
believe the point data is written in the order it is traversed in the
vtkPointData container.


On Mon, Dec 2, 2013 at 10:19 PM, Ben Colcord <ben.colcord at gmail.com> wrote:

> Maybe I misunderstand, but if the problem was because of point data arrays
> being different on different blocks, wouldn't the same errors occur in
> serial? I only get problems running in parallel.
>
>
> On Mon, Dec 2, 2013 at 8:20 PM, Andy Bauer <andy.bauer at kitware.com> wrote:
>
>> The point data arrays are probably ordered differently on separate blocks
>> or there are different point data arrays on different blocks. My suggestion
>> is to use a python programmable filter to remove the unneeded arrays or to
>> reorder them.
>>
>> Andy
>>
>> On Monday, December 2, 2013, Ben Colcord wrote:
>>
>>> Hello,
>>>
>>> I'm trying to extract data along a streamline. I have a script that
>>> works fine in serial, but gives errors in parallel. I get the same errors
>>> if I use the GUI or if I run the script with pvbatch, so I don't think it's
>>> a problem with the script. The errors I get are:
>>>
>>> ERROR: In /nv/hp25/bcolcord3/ParaView/VTK/Common/Core/vtkDataArray.cxx,
>>> line 474
>>> vtkDoubleArray (0xd4058a00): The number of components do not match the
>>> number requested: 3 != 1
>>>
>>> Warning: In
>>> /nv/hp25/bcolcord3/ParaView/VTK/Common/DataModel/vtkTable.cxx, line 421
>>> vtkTable (0xd4634a60): Cannot assign this variant type to
>>> multi-component data array.
>>>
>>> These errors occur when I do "writer.UpdatePipeline()" in the script, or
>>> "Save Data" in the GUI. Output is still produced, but the attached images
>>> show the problem. The streamline co-ordinates are OK, but the density along
>>> the streamline is not.
>>>
>>> I'm using ParaView 4.01, commit 8bab4fd99bda5b0b8748. My dataset is
>>> Multi-block, and I apply the following filters before attempting writing:
>>>
>>> CellDatatoPointData
>>> ProgrammableFilter (to remove ghost layers)
>>> Calculator (to create velocity vector from components)
>>> StreamTracer
>>>
>>> As you can see, it's a very simple task, but I can't get it to work in
>>> parallel. Any help would be greatly appreciated.
>>>
>>> Thank you.
>>>
>>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.paraview.org/pipermail/paraview/attachments/20131203/3f8dd976/attachment.htm>