Parameter scan_number_array is not updated during concatenation

[rlbyrne]

Related to closed issue #1159

Combining UVData objects fails when parameter scan_number_array is supplied. I am trying to concatenate two time-adjacent UVData objects read from .ms files. Each have 62128 blts, so the combined object should have Nblts=124256. The concatenation fails with error ValueError: UVParameter _scan_number_array is not expected shape. Parameter shape is (62128,), expected shape is (124256,).. Setting uv.scan_number_array=None for each object before combining bypasses the error.

[bhazelton]

Thanks for the report! Did you get this error when you were using the fast_concat method or just adding them with the + operator?

[rlbyrne]

Both!

[bhazelton]

I see the problem, we currently update all the parameters that are affected by name. We need to programmatically detect which parameters have a particular axis so when parameters get added everything just automatically works, like we've done for the select method. I've started a branch on this called add_concat_use_axis which fixes this but I want to extend it to the other objects as well, so it'll probably take a little while to get it in.

[rlbyrne]

Is it worthwhile to refer to this parameter by name or should we wait until the other fix is implemented?

[bhazelton]

Can you see if the branch I mentioned works for you? I did some testing and I think it does. It starts along the path of a general solution, so I think it's the right direction, but I want to get some feedback on it and to extend it to the other objects.

[bhazelton]

@rlbyrne sorry for the delay, but the PR linked above should fix this bug for you.

[rlbyrne]

Sorry I've been really slow to come back to this. I tried the add_concat_use_axis branch, but unfortunately it didn't resolve the issue.

[bhazelton]

Sorry I've been really slow to come back to this. I tried the add_concat_use_axis branch, but unfortunately it didn't resolve the issue.

Really? Fascinating. I added a test that I thought replicated what you were doing. It fails with the same error message as you report on main but it passes on the add_concat_use_axis branch, so I thought I was in the clear. Can you paste some code in here that shows what you're doing that still errors on that branch?

[rlbyrne]

Ok it looks like it is working now! The recent updates to the add_concat_use_axis branch seem to have fixed things.

[bhazelton]

Oh good. Thanks for checking!