Simple example:
XLS Reader (list of smiles strings) -> Molecular Type Cast -> RDkit Fingerprint ->
Column Rename (make all columns strings) -> CSV Writer.
Smiles column in csv file contains "Adapter for RDKitMolValue, StringValue, SdfValue, SmilesValue" instead of the smiles string...!
Is this a bug, or do I need to aproach this a different way?
Don't use the Column Rename node to "convert" data types. It does not do any conversion, instead it just forctes a new type into columns without touching the data. After the RDKit node the molecule column is not a Smiles column any more by a so-called adapter that contains several representations of the same molecule. This is why you get this rather strange output in the CSV file. If you want write Smiles, you need to convert the RDKit column to Smiles explicitly using the RDkit to Molecule node.
OK, thanks. I did notice that Write XLS node outputs the correct smiles string, so apparently that node manages to do the conversion? I would have thought that internal data types would all have a sensible "toString" method!
I've seen others using the Column Rename Node in this way - seems that if it works it's doing so by accident?!
To add to Thorsten's explanation: Yes, it is (resp. was) also a bug. This was fixed recently and is available in the Nightly build of the RDKit Nodes. You will get the SMILES value now also from an RDKit Adapter Cell instead of the strange text you encountered.
-Manuel