Has something changed in data trees behaviour?

Hi all,

I’m experiencing something weird with some components that do not behave as I thought they did for the past years when they are given data trees with different branches structures.

Note that I’m not searching for workarounds, I have components or ways to correct this. I’m just puzzled by the fact that this seems to suddenly happen and I would like to understand if it’s normal or not.

It seems the problem happens only when :

  • the output of the component is as list.
  • the component has at least two inputs
  • the tree structure has branches with different levels
    The outputs are then incorrectly grafted or flattened for no apparent reason. It’s like leading or trailing zeros are added and that causes some branches to merge where they should remain separated.

DataTreeProblem.gh (21.2 KB)

@magicteddy can you indicate in that image at each of the outputs what you would expect and (if different) what you would want as the outcome?


The blue panels show the tree structure I would expect instead of the red panels. Basically, that’s keeping the white panel structure, or adding a single {0} at the end, which I perfectly understand is necessary if a second input is as item.
Green panels are OK.

DataTreeProblem.gh (26.5 KB)

Do you compare it with an earlier build of Rhino 7? This is what I get in Rhino 5: ( I had to feed it a plane and there is no mid comp)
As far as I know and asked around there have been no changes to the core of gh

1 Like

I guess I should have tested an earlier version. At least this answers the question : it has always been there and it just seems I never encountered this case in all those years !

I still think that some results are strange (especially the Explode component that gives 4 branches instead of 2 with a complete mismatch of the edges), but as said above this was just out of curiosity. Probably better to leave it as is and not change the core behaviour.

Thanks for your time !