Looking to find duplicates using DIFFERENCE() among 2+ columns - sql

Question

I'm trying to write a SQL Select query that uses the DIFFERENCE() function to find similar names in a database to identify duplicates.
The short version of the code I'm using is:
SELECT *, DIFFERENCE(FirstName, LEAD(FirstName) OVER (ORDER BY SOUNDEX(FirstName))) d
WHERE d >= 3
The problem is my database has additional columns that include middle names and nicknames. So if I have a customer who has multiple names they go by, they might be in the database multiple times, and I need to compare a variety of columns against each other.
Sample Data:
+----+--------+--------+--------+--------+
|ID |First |Middle |AKA1 |AKA2 |
+----+--------+--------+--------+--------+
|1 |Sally |Ann |NULL |NULL |
|2 |Ann |NULL |NULL |NULL |
|3 |Sue |NULL |NULL |NULL |
|4 |Suzy |NULL |NULL |NULL |
|5 |Patricia|NULL |Trish |Patty |
|6 |Patty |NULL |Patricia|Trish |
|7 |Trish |NULL |Patty |Patricia|
+----+--------+--------+--------+--------+
In the above, rows 1+2 are duplicates of each other, as are 3+4, and 5+6+7.
So I'm not sure the best way to get what I want. Here's the longer version of the code I'm actually using:
WITH A AS (SELECT *,
SOUNDEX(FirstName) AS "FirstSoundex",
SOUNDEX(LastName) AS "LastSoundex",
LAG (SOUNDEX(FirstName)) OVER (ORDER BY SOUNDEX(FirstName)) AS "PreviousFirstSoundex",
LAG (SOUNDEX(LastName)) OVER (ORDER BY SOUNDEX(LastName)) AS "PreviousLastSoundex"
FROM Clients),
B AS (
SELECT *,
ISNULL(DIFFERENCE(FirstName, LEAD(FirstName) OVER (ORDER BY FirstSoundex)),0) AS "FirstScore",
ISNULL(DIFFERENCE(LastName, LEAD(LastName) OVER (ORDER BY LastSoundex)),0) AS "LastScore"
FROM A),
C AS (
SELECT *,
ISNULL(LAG (FirstScore) OVER (ORDER BY FirstSoundex),0) AS "PreviousFirstScore",
ISNULL(LAG (LastScore) OVER (ORDER BY LastSoundex),0) AS "PreviousLastScore"
FROM B
),
D AS (
SELECT *,
(CASE WHEN (PreviousFirstScore >=3 AND PreviousLastScore >=3) THEN (PreviousFirstSoundex + PreviousLastSoundex)
WHEN (FirstScore >= 3 AND LastScore >=3) THEN (FirstSoundex + LastSoundex)
END) AS "GroupName"
FROM C
WHERE ((PreviousFirstScore >=3 AND PreviousLastScore >=3) OR (FirstScore >= 3 AND LastScore >=3))
)
SELECT *,
LAG(GroupName) OVER (ORDER BY GroupName) AS "PreviousGroup",
LEAD(GroupName) OVER (ORDER BY GroupName) AS "NextGroup"
FROM D
WHERE (D.GroupName = D.PreviousGroup OR D.GroupName = D.NextGroup)
This lets me group together bundles of potential duplicates and it works well for me. However, I now want to add in a way to check against multiple columns, and I don't know how to do that.
I was thinking about creating a union, something like:
SELECT ClientID,
LastName,
FirstName AS "TempName"
FROM Clients
UNION
SELECT ClientID,
LastName,
MiddleName AS "TempName"
FROM Clients
WHERE MiddleName IS NOT NULL
...etc
But then my LAG() and LEAD() wouldn't work because I'd have multiple rows with the same ClientID. I don't want to identify a single Client as a duplicate of itself.
Anyways, any suggestions? Thanks in advance.