Chemistry: SDF reader

Not sure if that is in the right forum here, but couldn't find any better.

My problem regards the column handling in the SDF file reader: if I chose to NOT import some of the columns available in the SDF file, they are also not shown in the Knime tables subsequently. But apparently they are still present, since on writing the table back into an SDF these omitted column(s) will be present again.

Of course they can't be avoided from export since they are not visible within Knime. Anyone had that problem before?

Happy to send a workflow describing the problem.




The SD Reader leaves the molecules untouched and only stores the original string. Extracting certain attributes into columns does not change the molecules itself. Therefore, if you write the SD column back into a file you still get the original molecules with all attributes.

If you want to filter out attributes, you need to extract the Molfile structures (there is an option in the SD Reader) and subsequently select this column and some attributes in the SD Writer. Since the Molfile does not contain any attributes you will only get the selected attributes in the output files.

Now that makes sense!





I have the same problem with the output of the "External Tools (Labs)" node. I use an external problem that generate a SD file. In this file there is a property that is invisible in the output table but present somewhere because if I write the output table in a SD file I have the invisible property writen.

I would like to do some filtering based on this property. How could I do this? (except to write the SD file and re-read it again)




You can use the "SDF Extractor" node to extract properties from existing SDF molecules into columns.

Many thanks!

I will try to search more deeply in the existing nodes the next time!