Fuzzy Match Output

Hi Experts

I had this sample workflow downloaded to understand Fuzzy Match

The output is like below

Could someone please clarify why we have multiple rows for a given combination with same match score ? And the count differs - 3, 2, 1 etc ?

Examples - Record ID - 1 - count = 3 ; Record ID - 7 - count = 2 ; Record ID - 10 - count = 1

Thanks,

Umashankar

Fuzzy Match Sample.yxmd

Developer

Output

Fuzzy Match

Behavior Analysis

Accepted answers

All comments

shancmiralles

HI @umashankar_cns

Based on your sample:

RecordID Name Address

1	Annabel Gillingham	406 Campfire Street
20	Annabelle Gillingham	406 Camp Street

Fuzzy Match calculates Name & Address scores separately.

Each match gets recorded individually, even if they’re part of the same entity.

If both Name and Address pass the 80% threshold, two separate match records are created.

If a record matches itself (self-matching issue), it may add a third duplicate row.

so..

One row is for the Names match.

One row is for the Address match.

One row is for the overall combined score.

hope this helps!

and i was wondering if can share this publicly. This is a great sample data.

umashankar_cns

Thanks

But even if remove 'Address' in the fuzzy match and just match based on 'Name' alone, I am getting the same result

i.e., 3 , 2 or 1 rows in the result ?

shancmiralles

hi! @umashankar_cns

from how i see it, the fuzzy match decision is based on the fields you chose to match ( name & address).
if you take your record id#1 as a sample:

1	Annabel Gillingham	406 Campfire Street	Input A
20	Annabelle Gillingham	406 Camp Street	Input B

your configuration's output is:

1	20	92	96	87
1	20	92	96	87
1	20	92	96	87

now if we only have "Name" as the field on the Match Fields, your result is:

1	20	96	96
1	20	96	96
1	20	96	96

still three records, same as the original output.

Now let's try the address:

now we have one.

so it's the configuration for the "Names" that needs fixing.

Going back to the fuzzy match that only have "names" for the match"

You’re using DblMetaphone for Name Matching, Alteryx may be generating multiple phonetic keys for "Annabel Gillingham" and "Annabelle Gillingham". Each phonetic key can create a separate record in the match output.

I modified your settings for the match field ( Edit Match Option ) and changed "DblMetaphone " to none and and under match Functions I chose " Character ( No Spaces) : Levenshtein Distance:

1	20	90	90
2	17	87	87
3	12	93	93
4	21	100	100
5	16	85	85
6	13	90	90
7	14	94	94
8	19	92	92
9	22	94	94
10	18	93	93

comparing my output to your original N&A match fields your output did not provide IDs 4 & 9, while mine did not show 11.

if i apply my "Names" field configuration to your N&A it will show a lesser result.

So I configure the address field as well:

Generate Keys: None

Match Function: Words: Jaro Distance

Output:

1	20	89	90	87
2	17	90	87	94
3	12	92	93	92
4	21	99	100	97
5	16	88	85	92
6	13	90	90	90
7	14	97	94	100
8	19	93	92	94
9	22	96	94	98
10	18	96	93	100

Levenshtein Distance vs DblMetaphone: When to Use Each?
Use DblMetaphone when comparing names, addresses, and phonetic variations.
Use Levenshtein Distance when handling typos, spelling errors, and text-based differences.
For Best Accuracy: Combine both methods for strong fuzzy matching in Alteryx.

try doing a look up ( join tool ) to filter unmatched possible matches to make ID# 11 show.

I hope this somehow clear some stuff..

Thanks!

umashankar_cns

Thank you !

shancmiralles

@umashankar_cns

You are welcome!

I hope I gave you the resolution you need ..

I hope it okay I'll be borrowing your data case.. thanks!

Quick Links

This months top contributors

atcodedog05 19458

Qiu 15866

binu_acs 15708

MarqueeCrew 13708

apathetichell 13703