Community Spring Cleaning week is here! Join your fellow Maveryx in digging through your old posts and marking comments on them as solved. Learn more here!

Alteryx Designer Desktop Discussions

Find answers, ask questions, and share expertise about Alteryx Designer Desktop and Intelligence Suite.
SOLVED

Simple Fuzzy Match challenge

gnans19
11 - Bolide

Trying to fuzzy match below names

 

KUMAR, MR. JOHN T

Kumar, John

 

Please help me with Fuzzy Match Tool configuration to match the above names. Attached workflow.

11 REPLIES 11
Kenda
16 - Nebula
16 - Nebula

Hey @gnans19! My suggestion would be to use RegEx to get rid of any "Mr." or "Mrs." qualifications in your name field before performing the fuzzy match.

 

One other thing you could try would be to separate the first name and last name to separate fields and then fuzzy match on them both. Hope this helps!

gnans19
11 - Bolide

@Kenda Yes, I have already got rid of Mr. Mrs. Ms.and it worked

 

I noticed fuzzy match configurations hwich can ignore puntuations and salutations.

I would like to see this working using configurations chages.

Kenda
16 - Nebula
16 - Nebula

@gnans19 If you click on 'Edit' next to  you match style, there is an option that you can choose to strip punctuation and salutations. Is this what you're looking for?

 

Fuzzy Salutations.PNG

gnans19
11 - Bolide

Yes.. I tried this configuration change. But it didn't work. I would like to see a workflow with my sample values matched. I am interested in seeing the configurations.

MarqueeCrew
20 - Arcturus
20 - Arcturus

@gnans19,

 

If you use fuzzy matching and test with 1 case, you might be in for some unexpected results later on in the process.  I would caution against only matching with the Name field.  My preference would be to have an exact match on the ZIP or very tight match on the address components and a looser matching on the name field.  If name only matching is required, you can configure for that but do expect to have over-match results.  Even an exact match for John Doe can produce bad results when multiple John Doe's exist in the data.

 

just 2 more cents from the peanut gallery.

 

Cheers,

Mark

Alteryx ACE & Top Community Contributor

Chaos reigns within. Repent, reflect and restart. Order shall return.
Please Subscribe to my youTube channel.
MarqueeCrew
20 - Arcturus
20 - Arcturus

@gnans19,

 

The reason why you are not getting results is that while the Match Threshold is set to 70%, if you look within the EDIT of the Match Style, the "Name" style has a  Match Threshold of 85% set.  You can edit that to 75% and you'll see results matching (3 matches).  You would unique the results to get to your match.  I happen to generally use "Best of Jaro and Levenshtein" as my default.

 

Cheers,

Mark

 

 

Alteryx ACE & Top Community Contributor

Chaos reigns within. Repent, reflect and restart. Order shall return.
Please Subscribe to my youTube channel.
gnans19
11 - Bolide

Thanks @MarqueeCrew

I took this sample from my large dataset. I couldn't make it match even after playing around with Fuzzy tool configuration. Just wondering if someone can match and share the workflow.

MarqueeCrew
20 - Arcturus
20 - Arcturus

I made the edit described above.

 

Cheers,

Mark

Alteryx ACE & Top Community Contributor

Chaos reigns within. Repent, reflect and restart. Order shall return.
Please Subscribe to my youTube channel.
gnans19
11 - Bolide

@MarqueeCrew I ran the workflow, but there is no match

Labels