Cluster input removes empty branches and messes up tree structure

I have a data tree that has empty branches. When I feed it into a cluster, it gets confused and at the input parameter it will first try to clean out the empty branches. At the output parameter it seems to try to recreate these empty branches, but sometimes it does so erroneously. It is easy to try if you dispatch a tree and only feed one output into a cluster.

I now realize that if I"clean" the tree and "remove all empty branches", the cluster will work properly, but this will make the structure incompatible with other parallel tree structures.

Below is a data stream of 5 branches. I have dispatched it so the dispatched stream only contains 3 values, but still 5 branches. In this example the output of the cluster is as expected, but inside the cluster, the output shows something else.

I input these streams into a dummy cluster. Looking inside the cluster, the empty branches seem to have been "purged out", making for unexpected behavior inside the cluster:

I also have another example. The cluster input removes empty branches, but will not recreate them at the output:

This file is also attached.

Any thoughts on this one?

Attachments:

Cluster_Outside.gh, 11 KB
Cluster_Issue_02.gh, 70 KB

Replies to This Discussion

Permalink Reply by Andrew Heumann on October 19, 2014 at 10:08am

This is certainly a bug - and an annoying one at that. I typically rely on a short script that adds nulls to every empty branch - then if necessary I can clean the tree once inside the cluster.

private void RunScript(DataTree<object> x, ref object A) { foreach(GH_Path p in x.Paths){ if(x.Branch(p).Count == 0){ x.Add(null, p); } } A = x; }

Permalink Reply by Ramon van der Heijden on October 20, 2014 at 12:39am