SQL Server :Remove duplicate records but be able to choose which record to keep by specific column -

i have problem , can't find right solution. hope can me.

i have table looks this:

id  date      description     location 1  2012-12-01   low              1  2012-12-01   low              1  2012-12-01   low              b 2  2012-12-01   high             3  2012-12-01   low              c 

the result should this

id  date      description     location 1  2012-12-01   low              b 2  2012-12-01   high             3  2012-12-01   low              c 

i hope me.

a general method prioritize rows use row_number() case statement.

in case, seem be:

select t.* (select t.*,              row_number() on (partition id, description                                 order location desc                                ) seqnum       table t      ) t seqnum = 1;