My first serious project has hit the buffers and I wonder if i could please get some pointers to move forward. I need to form a set of groups dynamically based upon duplicate attribute values. Each group must consist only of duplicates or only a single primitive if there are no duplicates. Ideally I wish to store the result in another attribute but forming a group for each set of duplicates will suffice. How to get this done most efficiently? I need a “Group by Duplicate attribute” SOP or VEX wrangle equivalent but just cant quite get the code logic right as my screenshot shows.


Specifically in the below annotation I need the values listed in perimInt (right column) to be allocated a sequential mesh ID (2nd from right column) based upon descending order of size ( these relate to perimeter size of each primitive by colour) , so primitives 0-1 should be meshID 0, primitive 3 is meshID 1, primatives 3-4 is MeshID 2 etc

I have tried the partition SOP to create the groups but the rule syntax is problematic if you do not know in advance the attribute values you are grouping against. I am a VEX newbie so have tried to write my own wrangle but have underestimated the programming complexity of the task which i initially thought would be straightforward. Any helpful suggestions gratefully appreciated.

Never mind. managed to solve it in a Wrangle by just rethinking the logic. I was over thinking it. This simple code works for my attribute list provided it has been sorted into descending order by the preceding Sort node.

or you can just do this, including the sorting
Prim Wrangle:

tamte
or you can just do this, including the sorting
Prim Wrangle:

int vals[] = uniquevals(0, "prim", "perimInt");
vals = sort(vals);
vals = reverse(vals);
i@meshID = find(vals, i@perimInt); 

Elegance personified Tomas ,thank you, that is so much better and removes a node as well. Vex knowledge++