Each sd-file is loaded and parsed such that one output row of the table contains the data for 1 compound in the sd-file. The node attempts to parse the header block, subject to a number of caveats. Firstly, not all SDF writers produce correctly formatted MOL block headers, in particular in the second line. Incorrectly formatted lines will be parsed as if they were correctely formatted, which might result in spurious contents in some columns. The parsing is permissive, so the node will not fail to execute with poor formatting. Secondly, V3000 mol and sd-files use '0' counts in the header counts line. It is these counts which are reported, not the new 'M V30 COUNTS ...' line
This node was developed by Vernalis Research . For feedback and more information, please contact knime@vernalis.com