I have a collection full of objects transformed to form a structure. In GN, I'm basically taking the objects in as instances, resetting them and instancing the reset pieces onto a circle. What I want to be able to do after this is to be able to mix the original, unresetted positions of the instances to those that have been generated after distribution.
Despite my best efforts I was unable to do so. I tried everything like capturing the instance transforms of the original positions and mixing that with the captured positions after distribution, absolutely to no avail :( This is a breeze to do without a second thought in Houdini with packed geometry, yet in Blender I really can't seem to make sense of things in GN. The viewer node is able to display the world positions of both distributed and original instances perfectly, yet when it comes to reading them, zilch... :(
Any help would be wildly appreciated.

