select DISTINCT rowid, column1, column2 from mytable

Hi,
if you would like to distinguish between two identical rows you have to
change your query to use some row numbering. SQL 2005 provides you mechanism
to do that:

select
 NumberedTable.*
from
(
 select
   column1, column2,
   ROW_NUMBER() over (order by column1, column2) as RowNum
 from myTable
) NumberedTable

If you do not use SQL 2005 you have to create temporary table with one more
identity column and first fill this table with your date and after that
select that data from this table.

There is still question if your identical rows will always be in same order
but who knows? :)

Regards,
Ladislav