Like?? Like won't find similarities in strings with transposed, extra or omitted characters. Damareau-Levenshtein specifically targets what typically happens in typos.I kind of disagree in a db because you can probably build a much better search using sql Like search. This is based on a pretty thorough demonstration of a couple of interfaces using (Soundex, Simil, Levenshtein, and just find as you type)
The database in your attachment looks interesting. Nice to have all those techniques combined in one search. I'll definitely be taking a good look at it. Always on the lookout for ways of finding similarities in big databases. Thanks.
Clearly the Levenshtein Distance is an important component of your search. Note that Damareau adds detection of transposition, omissions and insertions which are not covered by Levenshtein but is considerably more expensive to process. I think D-L covers similar ground to Simil too.
I eventually converted the D-L function to VB.NET and have it as a CLR function in SQL Server, which overcomes the performance limitations. Someone showed me a native TSQL version of it too. I compared the performance and found the SQL version was sixty percent slower than the CLR.