T-SQL: Parsing names to ignore spaces and middle initials - sql

Question

I have a poorly maintained database that includes employee information. Human Resources requested a report that lists instances where the employee name associated with an insurance coverage does not match the name on the insurance policy.
There are inconsistencies in the formatting of the names in both tables. It's always last name then first name, but you might see any of the following in either table for a fictional employee named Steven J. Smith:
Smith, Steven
Smith,Steven
Smith, Steven J.
Smith,Steven J.
I need to run a query looking for instances where EMPLOYEE.EMP_NAME <> INSURANCE.SUBSCRIBER_NAME while allowing for differences in name formatting as shown above (i.e. picking up that "Smith,Steven J." and "Smith, Steven" are (probably) the same person and igonring them).
SELECT
EMPLOYEE.EMP_NO
, EMPLOYEE.EMP_NAME
, INSURANCE.SUBSCRIBER_NAME
, INSURANCE.PAYOR_NAME
FROM EMPLOYEE
INNER JOIN INSURANCE ON EMPLOYEE.EMP_NO = INSURANCE.EMP_NO
WHERE EMPLOYEE.EMP_NAME <> INSURANCE.SUBSCRIBER_NAME
I know I want to do a substring to ignore the middle initial, but how do I account for ignoring whether or not there is a space after the comma?

Answer 1

Why not just remove all commas and spaces with REPLACE?
WHERE REPLACE(REPLACE(EMPLOYEE.EMP_NAME,' ',''),',','') <> REPLACE(REPLACE(INSURANCE.SUBSCRIBER_NAME,' ',''),',','')

Answer 2

You could simply replace out the comma
WHERE replace (EMPLOYEE.EMP_NAME,',','') <> replace (INSURANCE.SUBSCRIBER_NAME,',','')
To find most mismatches...
;with cE as
(select
EMP_NO,
REPLACE(REPLACE(REPLACE(EMP_NAME,',',''),' ',''),'.','') as namekey
from EMPLOYEE),
ci as
(select
EMP_NO,
REPLACE(REPLACE(REPLACE(SUBSCRIBER_NAME,',',''),' ',''),'.','') as namekey
from INSURANCE)
select *
from ce
inner join ci on ce.EMP_NO = ci.EMP_NO
where
not
(
(LEN(ce.namekey)< LEN(ci.namekey) and ci.namekey like ce.namekey+'%')
or
(LEN(ce.namekey)>= LEN(ci.namekey) and ce.namekey like ci.namekey+'%')
)

Answer 3

you can remove space after comma and then remove initials
declare #Temp table (Name nvarchar(128))
insert into #Temp
select 'Smith, Steven' union all
select 'Smith,Steven' union all
select 'Smith, Steven J.' union all
select 'Smith,Steven J.'
select
case
when N1.Name like '% %' then left(N1.Name, charindex(' ', N1.Name))
else N1.Name
end as Name_New,
T.Name
from #Temp as T
outer apply (select replace(T.Name, ', ', ',') as Name) as N1

Answer 4

Thanks, your answers helped a lot. I ended up cutting the name into [lastname][firstname] with no spaces and cutting off the middle initial if it was there. Here's what eventually worked in eliminating the vast majority of the same-name matches:
((CASE
WHEN CHARINDEX(' ',REPLACE(REPLACE(EMPLOYEE.EMP_NAME,', ',''),',','')) = 0
THEN UPPER(REPLACE(REPLACE(EMPLOYEE.EMP_NAME,', ',''),',',''))
ELSE UPPER(LEFT(REPLACE(REPLACE(EMPLOYEE.EMP_NAME,', ',''),',',''),CHARINDEX(' ',REPLACE(REPLACE(EMPLOYEE.EMP_NAME,', ',''),',',''))))
END) <>
(CASE
WHEN CHARINDEX(' ',REPLACE(REPLACE(INSURANCE.SUBSCRIBER_NAME
,', ',''),',','')) = 0
THEN UPPER(REPLACE(REPLACE(INSURANCE.SUBSCRIBER_NAME
,', ',''),',',''))
ELSE UPPER(LEFT(REPLACE(REPLACE(INSURANCE.SUBSCRIBER_NAME
,', ',''),',',''),CHARINDEX(' ',REPLACE(REPLACE(INSURANCE.SUBSCRIBER_NAME
,', ',''),',',''))))
END))