Question

Possible Duplicate:
delete duplicate records in SQL Server

I have a table in which unique records are denoted by a composite key, such as (COL_A, COL_B).

I have checked and confirmed that I have duplicate rows in my table by using the following query:

select COL_A, COL_B, COUNT(*)
from MY_TABLE
group by COL_A, COL_B
having count(*) > 1
order by count(*) desc

Now, I would like to remove all duplicate records but keep only one.

Could someone please shed some light on how to achieve this with 2 columns?

EDIT:
Assume the table only has COL_A and COL_B

Answer 1

1st solution,
It is flexible, because you can add more columns than COL_A and COL_B :

-- create table with identity filed
-- using idenity we can decide which row we can delete
create table MY_TABLE_COPY
(
  id int identity,
  COL_A varchar(30), 
  COL_B  varchar(30)
  /*
     other columns
  */
)
go
-- copy data
insert into MY_TABLE_COPY (COL_A,COL_B/*other columns*/)
select COL_A, COL_B /*other columns*/
from MY_TABLE
group by COL_A, COL_B
having count(*) > 1
-- delete data from  MY_TABLE
-- only duplicates (!)
delete MY_TABLE
from MY_TABLE_COPY c, MY_TABLE t
where c.COL_A=t.COL_A 
and c.COL_B=t.COL_B
go
-- copy data without duplicates
insert into MY_TABLE (COL_A, COL_B /*other columns*/)
select t.COL_A, t.COL_B /*other columns*/
from MY_TABLE_COPY t
where t.id = (
               select max(id) 
               from MY_TABLE_COPY c  
               where t.COL_A = c.COL_A
               and  t.COL_B = c.COL_B
             ) 
go

2nd solution
If you have really two columns in MY_TABLE you can use:

-- create table and copy data
select distinct COL_A, COL_B
into MY_TABLE_COPY
from MY_TABLE
-- delete data from  MY_TABLE 
-- only duplicates (!)
delete MY_TABLE
from MY_TABLE_COPY c, MY_TABLE t
where c.COL_A=t.COL_A 
and c.COL_B=t.COL_B
go
-- copy data without duplicates
insert into MY_TABLE
select t.COL_A, t.COL_B
from MY_TABLE_COPY t
go