how to paint by file in parallel

Using a multiple file dataset, in parallel, I need to create a variable that has the file number in it. Is this possible, and how do I do it?

My use case is that I am trying to debug a filter, and am wondering if the issue is processes that don’t have any data. Thus, I want to use the clip filter to force some cores to be empty.

In general, I don’t think there is any way to identify from which file each cell comes from. That information would have to be produced by the reader, so would have to be implemented per format for each reader.

However, I’m not sure why you want that based on your use case. It sounds more like you want to know on which process that ParaView is running where the data are. In that case, you can just use the Process Id Scalars filter to get that information. From there applying the clip filter to make one process with no data should be straightforward.

Yep. I will write a feature request.

You are actually correct, my use case is not the best. A better example is when I have a few thousand (or tens of thousands) of files, I know one has an issue, but don’t know which that is. Doing a binary search of all of the files (especially since you can’t read a subset of the files in a dataset) is a pain.

Being able to identify the file would definitely be the best case. But in the mean time, if you are dealing with Exodus files, you might speed things up by identifying the global node or element ID, and then using SEACAS tools (maybe exotxt) to find that id in the files.

Of course, the big win would be if we could load individual file/files in exodus. Then, you could do a binary search. I have a bug in, but it’s complicated enough that the plan is to implement it correctly when w rewrite the Exodus reader.

The Exodus file reader still has the File Pattern options. If you want to do a binary search on the files, why not play with the File Range option?