Here is a quickie way i have done something similiar:
add a bit column to your table, like dupe_hold.
run a script that set the dupe_hold field to 1 based on the duplicates that you find.
Update TableName
Set dupe_field = 1
from TableName
where
(select count(UserName), UserName
from TableName
group by UserName
having count(UserName) > 1)
the run a script that grabs only one instance of the duped records and set that field to 0, zero.
Update TableName
Set dupe_field = 0
from TableName
where
(Select top 1
from TableName
where dupe_field = 1)
now delete the remaining records that have a 1 in the dupe_field column.
delete
from TableName
where dupe_field = 1
drop the field from the table and now your dupes are gone.