SQL Server: How can I use the COUNT clause without GROUPing?

Question

I\'m looking get two things from a query, given a set of contraints:

• The first match
• The total number of matches

I can get the first ma

Answer 1

Assuming you are using a newish version of SQL Server (2008+ from memory) then you can use analytic functions.

Simplifying things somewhat, they are a way of way of doing an aggregate over a set of data instead of a group - an extension on basic aggregates.

Instead of this:

SELECT
     ... , 
    COUNT(*) as MatchCount
FROM Table
WHERE
     ...

You do this:

SELECT
     ... ,
    COUNT(*) as MatchCount OVER (PARTITION BY <group fields> ORDER BY <order fields> )
FROM Table
WHERE
     ...
GROUP BY

Without actually running some code, I can't recall exactly which aggregates that you can't use in this fashion. Count is fine though.

Answer 2

Try next query:

select top 1
    *, count(*) over () rowsCount
from
    (
        select
            *, dense_rank() over (order by ValueForOrder) n
        from
            myTable
    ) t
where
    n = 1

Answer 3

Well, you can use OVER clause, which is an window function.

SELECT  TOP (1)
    OrderID, CustID, EmpID,
    COUNT(*) OVER() AS MatchCount
FROM    Sales.Orders
WHERE   OrderID % 2 = 1
ORDER BY OrderID DESC