Comparing Options / Fuzzy Match Option

[Chad_Cox]

I have two files that I have already merged based on Commodity Code and removed the rows that have an exact match for the descriptions.  Now, I'm left with 6,000+ records and want to see if Fuzzy Match can give me a score on how close the descriptions match.  Is this an option?  Also open to other suggestions.  Sample data file is attached and in the attached image I want to know who close the two columns in red match each other.

[jdminton · Accepted solution]

Yes, you can do this. You will need to:

1. Start with 2 separate files (or separate the columns into 2 separate with two select tools)
2. Add recordIDs to each and make them unique (you can add text for the DOAS and UGA in front or make the RecordID on one table start with a number like 1000001.
3. Add filename or some kind of identifier for each set of data
4. Union the tables so the descriptions and record IDs are in the same columns. 
5. Feed them into a Fuzzy Match tool and select "Merge"
6. I customized the Match Style on the description column. You need to start Alteryx as Admin to do this. You can also try some of the standard, but I like to play with things.
7. Group the results so if one record generates keys that match the 2 set of records more than once, you eliminate the lower value.
8. Join back to the first set on recordID
9. Join back to the second set on recordID2.
10. Sort by record ID and record ID2 ascending