I'm glad my earlier answer was useful to you. You didn't like having to represent each of your possible attributes by one-character symbols...
Ok, as I try to signal to you in the comments there, it is possible to adapt the Levenshtein Distance algorithm to look not at each character in a string, but at each element of an array instead, and do comparisons based on that. In fact it's quite straightforward to make this change:
Before 'Step 1, convert your comma-separated strings into arrays like this:
Dim sSplit() As String
Dim tSplit() As String
sSplit = Split(s, ",")
tSplit = Split(t, ",")
Then replace these four lines of code
N = Len(s)
m = Len(t)
s_i = Mid$(s, i, 1)
t_j = Mid$(t, j, 1)
with these
N = UBound(sSplit) + 1
m = UBound(tSplit) + 1
s_i = sSplit(i - 1)
t_j = tSplit(j - 1)
The + 1 and - 1 are there because Split returns a zero-based array.
Example usage:
?LevenshteinDistance("valros,helmet,42","valros,helmet,42")
0
?LevenshteinDistance("valros,helmet,42","knight,helmet")
2
?LevenshteinDistance("helmet,iron,knight","plain,helmet")
3
Note that 0 means the two strings are identical. You don't need separate code to deal with this.
With the above you should be able to complete your task.
One more note: the Damerau–Levenshtein distance may be a more relevant algorithm for you than the Levenshtein distance. The difference is that in addition to insertion/deletion/substitution, the D-M distance also considers transposition of two adjacent characters. Up to you to decide.
