unable to change the source file using python

ParaView Support
1

Hi everyone,

Currently using the version 5.5.0, I tried modifying the filename of the source file in my pipeline. However, updating the pipeline does not seem to trigger the source to re-read the data. This was working fine with 5.3.0.

How should I now change the filename of my source using a python script, having to load all files like it was a time-series.

This is done using the CGNSSeriesReader.

Thanks in advance

2

To provide more details, in version 5.3.0, I was using this kind of python code to update a pipeline and get data from it:

for i in range(10):
    print i
    _CGNSFileName = 'path/to/my/data/solution_{0}.cgns'.format(i)
    # Properties modified on out00000600_solutioncgns
    inputCGNS.FileNames = [_CGNSFileName]
    inputCGNS.UpdatePipeline()
    # Post-processing / other filter here
    RenderAllViews()

In version 5.5.0, this method does not update to data. I do not want to load all my files at once and then animate time.
Is there a solution to this ?

Thanks in advance

3

That does sound like a bug. Do you have easy access to a couple of timesteps that you can share to reproduce this? If not, I am sure I can locate a dataset too.

4

Hi,

Unfortunately, I do not have a small enough dataset to send …

In the meantime, I found an other solution: I change the input data of my “top level” filters to new source CGNS file. this is working. So I am not sure what is the best in terms of memory management between my first approach (changing the filename), and the last one (reading the new data, change the filter input, remove the old source)

Pierre

5

The bug is still present with ParaView 5.7.
Last time changing the FileNames was working correctly was with version 5.4.1 of ParaView.
It seems that the CGNSSeriesReader may not be transferring the new FileName to the CGNSReader (that is used underneath for one file only) when doing the loop.
Right now the only way is to delete the object and create a new CGNSSeriesReader every time.