I have a product classification rule which is set up as a negative look ahead regex; '^(?!RT-|OW-|MC-|NA|[0-9]+).*' So if the product name does not meet that criteria it is then added to that classification report. I believe its because I am doing this I get an extraordinary number of 'unpecified' values in that particular product classification report (80%-90% of our products would meet that Regex criteria and thus is excluded). I cannot build a rule based on the 'included' product names as they are very dynamic and would be a maintenance headache so I thought this is the best work around.
So I was wondering if importing a classification data file with a rule to set all those unspecified to a generic label might work? I don't think I can use the ~none~ key or is that incorrect? Or is there a ~unspecified~ key I could use in the import file?
Ah got it, totally forgot about the rules processing priority and how the last matching rule takes precedence. I modified my product rules such that now for each set classification the first rule for each set classification is the “wild card” mapping to “Others”. Then each subsequent rule for within each set classification classifies what it should be in the product name accordingly for that specific set classification.
I have published the new rules and will let you know in a couple of days if that resolved the problem but it makes sense it should result in zero unspecified. Hopefully my “others” in each will be very minimal.