Question

I have written this query:

SELECT s, [1] AS a1, [2] AS a2, [3] AS a3, [4] AS a4
FROM (SELECT grade, aid, s FROM m) p
PIVOT
(
SUM(grade)
FOR aid IN ([1], [2], [3], [4])
) AS pvt ORDER BY pvt.s;

That returns the result:

s  a1  a2  a3  a4
1  25  69  95  56
2  27  99  16  87
. . . .
99 98  12  34  76

Which is exactly the result I want. My problem is that there will not always be four distinct values in ‘aid’. Is it possible to rewrite this query (or use a stored procedure) so that the amount of ‘a*’ columns depends on how many distinct values are in ‘aid’?

Answer 1

You will need to use a Dynamic Pivot to get the list of columns that you want. This will retrieve the list of columns first and then pivot that list. Something similar to this:

DECLARE @cols AS NVARCHAR(MAX),
    @query  AS NVARCHAR(MAX);

select @cols = STUFF((SELECT distinct ',' + QUOTENAME(aid) 
            FROM m 
            FOR XML PATH(''), TYPE
            ).value('.', 'NVARCHAR(MAX)') 
        ,1,1,'')

set @query = 'SELECT s, ' + @cols + ' from 
            (
                select grade, aid, s
                from m
           ) x
            pivot 
            (
                sum(grade)
                for aid in (' + @cols + ')
            ) p 
            ORDER BY p.s'

execute(@query)