Question

I work with SQL Server 2008. There is table with projects, their status and date:

ID    Status      Date
1     Gone        21-10-2011
2     Cancelled   01-10-2011
3     Won         02-10-2011
4     Won         07-10-2011
5     Gone        02-01-2010
6     Won         24-01-2010

For statistics purposes I need some T-SQL query that in result will return table like this:

Month    Year    Total    Lost   Won
10       2011    4        2      2
01       2010    2        1      1   

Where:

Total means total entries amount for unique Month-Year pair.

Lost means total entries amount for unique Month-Year pair with status ‘Gone’ or ‘Cancelled’

Won means total entries amount for unique Month-Year pair with status ‘Won’.

I understand how can I get total amount for unique Month-Year, but don’t know how to add in result set Lost and Won columns and get entries amount. And of course if there are no any ‘Won’ projects for Month-Year pair and only ‘Lost’ it should be 0 value for Won column.

Answer 1

SELECT
   MONTH([Date]) [Month],
   YEAR([Date]) [Year],
   COUNT(*) [Total],
   COUNT(CASE WHEN Status IN ('Gone', 'Cancelled') THEN 1 END) [Lost],
   COUNT(CASE WHEN Status= 'Won' THEN 1 END) [Won]
FROM
   MyTable
GROUP BY
   MONTH([Date]),
   YEAR([Date])

This works because the silent ELSE in the CASE gives NULL, and COUNT will ignore this