Forum Moderators: coopster
similar_text();
soundex();
metaphone();
levenshtein();
Since levenshtein is O(m*n) and to similar_text() is O(max(n,m)**3), I would go with levenshtein if it works for you.
PS O (aka "Big O" in computer speak) is how an operation scales in terms of actions (rather than CPU time or memory or some physical measure). In other words, it's a mathematical measure of efficiency. So levenshtein takes m*n operations whereas similar_text takes the longest string, m or n, and then you cube it, so it scales very poorly to long strings.