Question

I have a table with votes from multiple sectors (think of each sector as a state or similar) on multiple candidates. Each sector has multiple candidates, each with a different vote count.

Here is my table (simplified)

CREATE TABLE [Results]
(
    [SectorID] BIGINT,
    [CanditateID] BIGINT, 
    [VoteCount] BIGINT,
    [Newness] DATETIME
)

I obviously keep sector and candidate meta data in another table, but I need to find the highest voted candidate for each sector, so I can join those tables into a view.

The best candidate in each sector, is determined by [VoteCount], and if there are two with the same number of votes, it is determined by [Newness]. There must remain exactly one line per sector, and I have to be able to use it in a view, joined together with the meta data.

How do I obtain the highest voted candidate from each sector?

Answer 1

Assuming that you are using SQL Server 2005 or greater, you want to do this with row_number():

select r.*
from (select r.*,
             row_number() over (partition by SectorId order by VoteCount desc, Newness desc) as seqnum
      from Results r
     ) r
where seqnum = 1