数据透视表Q-SQL Server 2008

Question

I have the following table ...

GROUPING    Num DEF  TOTAL   SOURCE
a           1   a    -2503   WTH
a           2   b    -180    DET
a           2   c    -156    PLY
a           4   d    -99     DET
a           5   e    -252    DET

Which I'd like to present as follows ...

GROUPING    Num    DET      PLY     WTH
a           1                       -2503
a           2      -180     -156
a           4      -99      
a           5      -252     

Here is my effort ...

select 
    grouping
    ,num 
    ,det
    ,ply
    ,wth
from #temp
pivot (
    sum([TOTAL]) for [SOURCE] in (det,ply,wth)
) x

But I get 2 rows with num = 2 where I want 1 row.

Any help appreciated.

Thanks, James

Answer 1

Your existing query is not using a subquery to select from your table, as a result all of the columns not used by the PIVOT are used in the GROUP BY . Since you have multiple distinct values for the DEF column you are returning multiple rows for num=2 .

I would use a subquery to select only the columns you need for the PIVOT and the final select list:

select 
    grouping
    ,num 
    ,det
    ,ply
    ,wth
from
(
  select grouping, num, total, source
  from #temp
) d
pivot 
(
  sum([TOTAL]) 
  for [SOURCE] in (det,ply,wth)
) x;

See SQL Fiddle with Demo