String Similarity

Two spellings of the same thing, or two different things? Five measures of how far apart they are, so you can pick the one that suits your data and see where to put the threshold.

The two strings

First

Second

Before comparing Ignore upper and lower case Trim and collapse spaces Ignore punctuation

How close

Each measure

Levenshtein counts single character edits. Damerau also counts a swap of two neighbors as one edit. Jaro-Winkler leans on a shared start, which suits names. Dice compares pairs of letters and cares little about order. Longest common run is what a diff would show you.

The edits between them

kept · changed · added · removed

Match against a list ranked against the first string above

Candidates

Call it a match above