Showing results for 
Search instead for 
Did you mean: 

Alteryx designer Discussions

Find answers, ask questions, and share expertise about Alteryx Designer.

The highly anticipated Alteryx Community tradition is back! We hope you'll join us!

Learn More

Fuzzy matching against all known info associated with an individual?


I'm trying to create an ongoing, regularly run, fuzzy matching module that would compare newly submitted individuals against a "roladex" of all known information for individuals seen before. I have both a database that has every unique entry (e.g. assigned ID Number Name, DOB, address, phone number, sex), one entry per line. This is what I've been using to match against with all new incoming records. However, I also have a database that is structured vertically (i.e. assigned ID number, data field name, value) that contains every unique element associated with that ID number. I'd like to be able to run fuzzy matching against this, but don't know if it's possible. After all, let's say Abbie Smith, 1/15/90. 123 Anywhere Ave, 555-555-5555, F moved also has a record at Abbie Smith, 1/15/90. 567 Newtown St, 555-666-7777, F. Then she comes in again, with a record that says Abbie Smith, 1/15/90. 567 Newtown St, 555-555-5555, F (maybe the first number is her cell phone). With a traditional fuzzy match, it may not pick up that this should be the same person, because it compares each record to each other record. But if instead I can match the new record against all known phone numbers, addresses, etc. for Abbie, using my roladex database, it would show that this is a valid combination. I have thought about creating fake records that contain every combination of every known element for every ID, but that could get unwieldy quite quickly. 


Any ideas...? 


Hello @AliKat I am not sure this is exactly what you are looking for, but you are trying to find an instance in a vertical record for the incoming value, I would suggest you to try "Find And Replace" tool where it will find if the word exactly matches or the column contains the incoming data type. Once you identify such records, then you can use it for your next set of action or process.


Please let me know if this is what you are looking for, if not happy to help.


No, because i'm not looking for an exact match. I'd like to use fuzzy matching. And I don't want to match against just one item, it would have to fuzzy match for the person's name, DOB and sex and then look at all addresses and phone numbers (the things that change) already known to be associated  with that person, and use any of those to fuzzy match to the incoming record. Does that make sense?