Delete duplicates

delete from a
from
dbo.tableWithDupes a
join (
select
ColPrimaryKey
,row_number() over(
partition by AccountId, Col2, Col3 -- these are the unique columns
order by AccountId -- use this to give one record priority over another
) DupeCount
from
dbo.tableWithDupes
) q
on a.ColPrimaryKey = q.ColPrimaryKey
where
q.DupeCount > 1