Handling of missing cells in neighbour grams


the parallel universes and neighbour grams approach is one of my favourite analysis techniques in knime. However I do find the handling of missing cells causes me to tear my hair out!

In pharma, very rarely do you have all data on all molecules. There are always gaps (I.e. missing cells) in different columns. I find it very useful to look for Nearest neighbours in the neighbour grams viewer but any molecules with missing data are automatically filtered out, and there is no other option on how to handle this. It would be useful if you could have an option for the molecules with missing values to be included in which that property distance is simply not calculated (I.e. missing value is equal to the value for comparison, or even better is maybe the distance used is the average difference from the other property distances calculated). It's still not a perfect solution, but would be better than simply removing the row.