Finding Similar Type of Names | Fuzzy Search in SQL

Fuzzy Search :
A technique of finding the strings that match a pattern approximately (rather than exactly). Users / Reviewers often capture names inaccurately. Typing mistake is a very common mistake that reviewer does while capturing the names which leads to inconsistency in Data. But sometimes, we need to search or match this inaccurate data anyway. For example, users should match existing customer records rather than creating unwanted duplicates.

Problem Statement :
Finding all kind of error types that described in the below table e.g. Spelling Mistake, Incomplete Name, Formatting Issue, Suffix Missing.

FirstName	LastName	FirstName	LastName	Error Type
Benjamin	Sheriff	Benjamin	Sherrif	Spelling Mistake
Cocoa	Fairchil	Cocoa	Fairchild	Spelling Mistake
Danuzio	Pinheiro	Danuzio	Pinherio	Spelling Mistake
Jeff	Voight	Jeff	Voigt	Spelling Mistake
Marie	Prendergast	Marie	Prendergrast	Spelling Mistake
Michael	Heinrich	Michael	Heinricy	Spelling Mistake
Nghia	Le	Nghia	Lee	Spelling Mistake
Leonardo	Madrigal	Leonardo	Madrigal Del Valle	Incomplete Name
Yolanda	Balcazar	Yolanda	Balcazar-Rodriguez	Incomplete Name
J	Garcia Martinez	J	Garcia-Martinez	Formatting Issue
Lauren	Lord – Tuley	Lauren	Lord-Tuley	Formatting Issue
Virginia	Pulley Alberts	Virginia	Pulley-Alberts	Formatting Issue
Connie	Marks	Connie	Marks Iv	Suffix Missing
Daniel	Ladd	Daniel	Ladd Jr	Suffix Missing
Rodrigo	Delgado	Rodrigo	Delgado Jr	Suffix Missing
Thomas	Griffin	Thomas	Griffin Iii	Suffix Missing
Wayne	Nance	Wayne	Nance Ii	Suffix Missing

Approach to solve :
SOUNDEX() Function can find the inconsistency in names. SOUNDEX() can evaluate the similarity of two names. SOUNDEX() only works well when we do have 1 or 2 tokens. Names generally contain 1-2 tokens, so it works fine with Name. But if we want to find similar names for the company then it’s not that useful as it contains multiple tokens.

Data :
Benjamin Sheriff and Benjamin Sherrif where FirstName is same but any of the LastName has a spelling error. Likewise, there are many errors in this data like formatting issues, incomplete names.

FirstName	LastName
Benjamin	Sheriff
Benjamin	Sherrif
Virginia	Pulley Alberts
Virginia	Pulley-Alberts
Geoffery	Bown
Geoffrey	Bown
Anthoni	Castoldi
Anthony	Castoldi
Rodrigo	Delgado
Rodrigo	Delgado Jr
Leonardo	Madrigal
Leonardo	Madrigal Del Valle
Mitchel	Myers
Mitchell	Myers

Code :
Case-1: FirstName is exact the same but LastName is a similar type.
Case-2: LastName is exact the same but FirstName is a similar type.

Let’s discuss it one by one as follows.

Case-1 : 
FirstName is exact the same but LastName is similar type :

select distinct ss.firstname,ss.lastname,sd.firstname,sd.lastname
from load as ss, load as sd
where ss.firstName=sd.firstName and SOUNDEX(ss.lastName)=SOUNDEX(sd.lastname)
      and
      left(ss.lastname,2)=left(sd.lastname,2)
      and
      ss.lastName<>sd.lastName

Output :

FirstName	LastName	FirstName	LastName
Benjamin	Sheriff	Benjamin	Sherrif
Benjamin	Sherrif	Benjamin	Sheriff
Leonardo	Madrigal	Leonardo	Madrigal Del Valle
Leonardo	Madrigal Del Valle	Leonardo	Madrigal
Rodrigo	Delgado	Rodrigo	Delgado Jr
Rodrigo	Delgado Jr	Rodrigo	Delgado
Virginia	Pulley Alberts	Virginia	Pulley-Alberts
Virginia	Pulley-Alberts	Virginia	Pulley Alberts

Case-2 :
LastName is exact same but FirstName is similar type :

select distinct ss.lastname,ss.firstname,sd.lastname,sd.firstname
from load as ss, load as sd
where ss.lastname=sd.lastname and SOUNDEX(ss.firstname)=SOUNDEX(sd.firstname)
      and
      left(ss.firstname,2)=left(sd.firstname,2)
      and
      ss.firstname<>sd.firstname

Output :

S.No.	FirstName	LastName	FirstName	LastName
1.	Anthoni	Castoldi	Anthony	Castoldi
2.	Anthony	Castoldi	Anthoni	Castoldi
3.	Geoffery	Bown	Geoffrey	Bown
4.	Geoffrey	Bown	Geoffery	Bown
5.	Mitchel	Myers	Mitchell	Myers
6.	Mitchell	Myers	Mitchel	Myers

Reference :
Github Link – https://github.com/SuryaSD/Finding-Similar-Types-of-Names-Fuzzy-Search