Fuzzy Match on ID combined with Char & Digits

Input:

I'm trying to get match with Merge mode of Fuzzy Match.

I tried with different "Generate Keys":

1.Double Metaphone w/Digits, -- Result: The key is based on Digits only so it give incorrect match as all Keyword contain digits 12345.

2.Whole Field ( Case Insensitive)-- Result: no match

3.Alphanumeric Only ( Case Insensitive) -- Result: no match

Alphanumeric Only MatchKeys:

I though fuzzy match will generate different length of keyword so i can pick possible match from output.

Need some help here as i can't figure out which "Generate Keys"/"match Function" to use in this case.

Fuzzy_Match_Test.yxmd

Workflow

Fuzzy Match

Iterative Macro

Accepted answers

ChrisTX

Take a look at the Match Function "Words & Digits: Levenshtein Distance"

https://blogs.mathworks.com/cleve/2017/08/14/levenshtein-edit-distance-between-strings/

Since your Keywords seem to be alphanumeric codes, the "edit distance" represented by Levenshtein should be a good option.

All comments

ChrisTX

Take a look at the Match Function "Words & Digits: Levenshtein Distance"

https://blogs.mathworks.com/cleve/2017/08/14/levenshtein-edit-distance-between-strings/

Since your Keywords seem to be alphanumeric codes, the "edit distance" represented by Levenshtein should be a good option.

BrandonB

What is the expected output? Could you give us some more information about the values that you are looking to match and why they wouldn’t be exact? The reason that I ask this is because if you have some strings that the first 5 characters are the same as an example it might be easier not to use fuzzy match at all. You could append the list of keywords to itself and then run it through a filter tool where left([keyword],5) = left([right keyword],5) and it would show you all of the keyword matches where the first 5 characters are the same without even needing a fuzzy match.

Corp-Yongcan-Frank-Lu

thanks.

The result matched expectation for these sample IDs.

1<-->1001

2<-->1002

3<-->1003

I will run against live data set tomorrow.

Quick Links

This months top contributors

atcodedog05 19598

Qiu 15867

binu_acs 15708

MarqueeCrew 13708

apathetichell 13703