community
cancel
Showing results for 
Search instead for 
Did you mean: 

Alteryx designer Discussions

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

Fuzzy Match matchscore

Alteryx Certified Partner

Hi,

 

I used fuzzy match quite a lot recently, but there is one question that makes me confused.

When I output the results, there are two matchscores (MatchScore, MatchScore_Data), I want to ask what are their difference (The scores are usually the same).

Are they the cross matching scores of two values? like A->B and B->A

 

1.jpg

2.jpg

 

Thanks.

 

Kenneth

Alteryx Certified Partner
Alteryx Certified Partner

@kennethli,

 

You have a match score (overall) and you have a match_score for each fuzzy component of the score.  If you had a name, address, zip, then you would have four scores.

 

At least, that's what I think happens.

Alteryx ACE & Top Community Contributor

Chaos reigns within. Repent, reflect and reboot. Order shall return.
Bolide

Hi @kennethli,

 

My understanding is the same as @MarqueeCrew. Each field you include in the fuzzy match tool gets its own match threshold. In order for a pair of records to be counted as a match, they must match well enough to exceed each of the individual field thresholds and then also the overall threshold you set (you can see the threshold for each individual field by clicking the Edit button next to where you choose the match style for a field).

 

If you want to more comprehensively learn the Fuzzy Match tool, check out Alteryx's webinar collection here: http://www.alteryx.com/virtual-training

 

Highlighted
Alteryx Certified Partner

Hi @DultonM and @MarqueeCrew,

 

Haha, right, I tested and found a video related to using fuzzy match.

https://www.youtube.com/watch?v=35aNGzJ5kYY

 

The matchscore_data is supposed to be the score for threshold defined in match function of each matching fields.

So that's correct when you defined four fields with match functions available (i.e. not match key only), then you have 4 scores + 1 overall score. 

 

That's very useful. Thanks!!!

an example of using these three Alteryx tools, and others to deduplicate a dataset using fuzzy logic
Labels