A modified version of the RDKit Dictionary Substructure Filter (and molecule version) node which appends the count of each substructure match to the output table (perhaps also allowing ALL records to be passed through to the output table - might be easiest simply by allowing the minimum matches counter to be allowed to be set to '0' in the filter node!?). I have posted further suggestions for how this might work in the Indigo forum for their existing substructuer counter node (http://tech.knime.org/forum/indigo/substructure-match-counter-and-query-molecule-to-indigo-nodes-suggested-imporvements), but I think having a similar node in the RDKit set would be very useful.
Further, the ability to count reaction substructure matches from reaction SMARTS would be useful (particularly if there was support for loading up .rdf files!)
Thanks
Steve