On Fri, 30 Nov 2018 05:32:13 +0100 Greg Landrum <[email protected]> wrote:
> I don’t think you should be getting NaNs, and that’s worth looking into, > but they are technically numeric values : that’s the double value younger > when you divide by zero. NaNs are great because you can tell "invalid result" from 0, or some other number that's perfectly valid per se, just out of bounds in your particular corner case. -- Dmitri Maziuk <[email protected]> _______________________________________________ Rdkit-discuss mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/rdkit-discuss
