Thanks for the new Pair Extractor node in the Distance Matrix section. This significantly advances the Distance Matrix utility, and chemistrywise allows Similarity Matched Pairs analysis to be carried out with ease now, and is a really powerful toolset for the Medicinal Chemist. This is really great.
I do have one minor niggle, the movement of the DistanceMatrix nodes into the Misc category, and then some of these nodes (Hierarchical, k-Medoids) to now move into Mining/Clustering is very disjointed and confusing. I am certain new users, will never find these nodes in the clustering section that can be applied to the Distance Matrix columnsets. I would consider moving the DistanceMatrix section from Misc and put this into the clustering section too as one possible suggestion here. Or at the very least, duplicate the Hierarchical/k-Medoid nodes in both locations.
Thanks for the feedback, I quite like the pair extractor node as well. We are also still not happy with the current organization, but hope it is a step in the right direction. The idea of showing nodes in more than one category is interesting. I'll look into that further.
I agree with you, this node is very useful. However, I think it will be better if it could give the user the possibility of adding a second optional result column.
Let me explain my issue. I have hundreds of molecules for which the distances have been pairwise calculated. When I use this node, I can extract the pairs either using the id of the molecules, or the representation of the molecules. In the first case I do not see the molecules, in the other case I see the molecules but without the id it is a hard task to find them in my dataset.
Adding an optional second column would allow me to have for each pair the representation and the corresponding id.
Currently, as a workaround I can use 2 joiners but it may be time consuming due to memory.
Is this something you could add in a next release?