I've been using SQL hacks to force empty groups and the like but I'm tired of doing it, I want a presentation layer solution for a presentation layer problem. I'm using SQL/SSRS 2008 R2 and I'm encountering more and more situations like the situation below.
I've got column groups for dates and I've got a row group for job state (cancelled or not) and status. My problem comes in when I have a day with no jobs or a month without a cancellation of a certain type. My requirements state that the report layout must be fixed and these rows/columns must still exist but just state 0. To make matters worse, the number of distinct statuses for one job state (cancelled) is dynamic in itself and pulled from a table, although, the number of statuses for the other job state (not cancelled) is fixed.
My question is, can I form this report layout:
https://i.sstatic.net/0UNZU.jpg
From these datasets?:
https://i.sstatic.net/zsqNw.jpg
-within SSRS or am I stuck with just having to continue SQL trickery to force that layout?
I know the question is quite old by now, if not you, I hope this answer can help somebody else.
Based in what you wrote, what you need are Common Table Expressions and a pivot table with dynamic columns.
Here is the code:
--
-- We create the needed schemas and fill them with your data.
--
CREATE TABLE Jobs(Date DATE, State NVARCHAR(32), Status NVARCHAR(32));
CREATE TABLE CancellationTypes (State NVARCHAR(32), Reasons NVARCHAR(32));
--
-- Added the Status column so we can get the cancellation types before
-- the jobs. This way we can get the zeros you need.
--
INSERT INTO CancellationTypes VALUES('Cancelled', 'Weather');
INSERT INTO CancellationTypes VALUES('Cancelled', 'Equipment Failure');
INSERT INTO CancellationTypes VALUES('Cancelled', 'Staff Caused');
INSERT INTO CancellationTypes VALUES('Cancelled', 'Misc');
INSERT INTO CancellationTypes VALUES('Not Cancelled', 'Completed');
INSERT INTO CancellationTypes VALUES('Not Cancelled', 'Partially Completed');
--
-- No changes here...
--
INSERT INTO Jobs VALUES('20140301', 'Cancelled', 'Weather');
INSERT INTO Jobs VALUES('20140301', 'Cancelled', 'Weather');
INSERT INTO Jobs VALUES('20140301', 'Cancelled', 'Equipment Failure');
INSERT INTO Jobs VALUES('20140301', 'Not Cancelled', 'Completed');
INSERT INTO Jobs VALUES('20140302', 'Not Cancelled', 'Completed');
INSERT INTO Jobs VALUES('20140302', 'Not Cancelled', 'Partially Completed');
INSERT INTO Jobs VALUES('20140305', 'Cancelled', 'Weather');
INSERT INTO Jobs VALUES('20140305', 'Cancelled', 'Equipment Failure');
INSERT INTO Jobs VALUES('20140305', 'Cancelled', 'Staff Caused');
INSERT INTO Jobs VALUES('20140305', 'Not Cancelled', 'Completed');
INSERT INTO Jobs VALUES('20140305', 'Not Cancelled', 'Partially Completed');
DECLARE @Headers NVARCHAR(158), @Query NVARCHAR(4000);
--
-- With the CTE we can get our range of dates and use it to create the
-- columns for the final resultset.
--
DECLARE @StartDate AS DATE = '20140301', @EndDate DATE = '20140305';
WITH Dates AS (
SELECT @StartDate AS Day
UNION ALL
SELECT DATEADD(D, 1, Day) FROM Dates WHERE Day < @EndDate
)
SELECT @Headers = STUFF((SELECT DISTINCT '],[' + CONVERT(VARCHAR(10), Day, 101) FROM Dates ORDER BY '],[' + CONVERT(VARCHAR(10), Day, 101) FOR XML PATH('')), 1, 2, '') + ']';
SET @Query =
'SELECT State, Status, ' + @Headers + ' FROM
(
SELECT
CONVERT(VARCHAR(10), J.Date, 101) AS Date,
CT.State,
CT.Reasons AS Status
FROM CancellationTypes AS CT
-- This can give us cancellation types with 0 jobs.
LEFT JOIN Jobs AS J ON CT.Reasons = J.Status AND CT.State = J.State
) t
PIVOT (COUNT(Date) FOR Date IN (' + @Headers + ')) AS pvt ORDER BY pvt.State'
EXECUTE (@Query);
--
-- We're done here...
--
DROP TABLE Jobs;
DROP TABLE CancellationTypes;
... and the results:
State Status 03/01/2014 03/02/2014 03/03/2014 03/04/2014 03/05/2014
------------- ------------------- ----------- ----------- ----------- ----------- -----------
Cancelled Equipment Failure 1 0 0 0 1
Cancelled Misc 0 0 0 0 0
Cancelled Staff Caused 0 0 0 0 1
Cancelled Weather 2 0 0 0 1
Not Cancelled Completed 1 1 0 0 1
Not Cancelled Partially Completed 0 1 0 0 1
Hope this helps.