Help needed: Fuzzy Match

[pvi]

Dear community, 

after hours of try-and-error and research, I have to declare defeat. 

wanted to create simple groups to be used in a tree. 

however, I still can't believe that such an easy task can turn out to be so tricky. 

thank you in advance for helping with this and for dealing with me as beginner-fuzzy-match-user.

[Jonathan-Sherman]

Hi @pvi,

A little confused as to what you're trying to achieve, are you aiming to get Column2 as the output from Column1?

Kind regards,

Jonathan

[Maskell_Rascal]

Is there a question here or are you just venting? If you'd like help please describe the problem, provide some sample data, and the expected output. 

[pvi]

apologies, deleted too much. 

my goal is to get from Column1 to Column2.

[atcodedog05]

Hi @pvi 

Is the groups predefined or should it be calculated?

[danilang · Accepted solution]

Hi @pvi 

I think you might be using the wrong hammer to drive this particular nail.  Fuzzy matching is used to find the "distance" between similar strings.   For example 123 Main Street, Anytown should match closely with 123 Main St., Anytown, but not with 123 Other Ave, Anothertown, even though all three strings start with 123.  Your case is similar in that 3M Libor Index TR CHF won't match closely with Libor, even though both contain Libor.   What you need is a strategy that can find keywords in a string.  

This workflow use the Find Replace tool to search for keywords anywhere in your input strings and append a column that contains the matched keyword.  The keywords here are the indexes providers that you're looking for, Libor, MSCI and I added a third AMERIBOR to illustrate the concept.  After the match, the results look like this

Dan

[pvi]

thank you very much for your help and your explanations.