Filter out aggregated rows SSRS - sql

Question

I have a fairly complicated report that I'm trying to edit. Below is a screenshot of the report:
I need to filter out all rows where "Number not in Groups" is n/a or 0, I've tried filtering out using the filter option in tablix properties, but it wont let me due to the error:
aggregate functions cannot be used in dataset filters.
Below is the code for the expression that makes up the field "Number not in Groups":
=iif(sum(iif(left(Fields!crs_group.Value,1) = "G",1,0),"GrpCourse")=0,
"n/a",
sum(iif(Fields!crs_group.Value="Enrolled on Course",1,0)) - sum(iif(left(Fields!crs_group.Value,1) = "G",1,0)))
I have also tried to filter out those rows in the SQL query but haven't had much luck. Below is my attempt at filtering out the rows in the query:
SELECT
sub.course ,
sub.crs_group ,
m.m_reference ,
me.e_status ,
me.e_id ,
s.s_studentreference ,
p.p_forenames ,
p.p_surname ,
pcd.p_surname + ',' + pcd.p_forenames AS course_dir ,
ISNULL(COUNT(*) - SUM(CASE WHEN LEFT(crs_group, 1) = 'G' THEN 1
END), 1) AS NumberNotInGroups
FROM
msql.unitesnapshot.dbo.capd_moduleenrolment AS me
INNER JOIN msql.unitesnapshot.dbo.capd_module AS m
ON me.e_module = m.m_id
LEFT JOIN msql.unitesnapshot.dbo.capd_staff scd
ON m.m_modulesupervisor = scd.s_id
LEFT JOIN msql.unitesnapshot.dbo.capd_person pcd
ON scd.s_id = pcd.p_id
INNER JOIN msql.unitesnapshot.dbo.capd_student s
ON me.e_student = s.s_id
INNER JOIN msql.unitesnapshot.dbo.capd_person p
ON s.s_id = p.p_id
INNER JOIN (
SELECT
m.m_id ,
CASE WHEN m.m_reference NOT LIKE '%G_'
THEN m.m_reference
ELSE LEFT(m.m_reference,
CHARINDEX('G', m.m_reference) - 1)
END AS course ,
CASE WHEN m.m_reference NOT LIKE '%G_'
THEN 'Enrolled on Course'
ELSE RIGHT(m.m_reference, 2)
END AS crs_group
FROM
unitesnapshot.dbo.capd_module m
) sub
ON sub.m_id = me.e_module
WHERE
me.e_status = 'A'
AND LEFT(m.m_reference, 2) = '12'
AND SUBSTRING(m.m_reference, 7, 2) IN ( 'VF', 'AB', 'FB' )
GROUP BY
sub.course ,
sub.crs_group ,
m.m_reference ,
me.e_status ,
me.e_id ,
s.s_studentreference ,
p.p_forenames ,
p.p_surname ,
pcd.p_surname + ',' + pcd.p_forenames
HAVING
COUNT(*) - SUM(CASE WHEN LEFT(crs_group, 1) = 'G' THEN 1
ELSE 0
END) <> 0
ORDER BY
p.p_surname ,
p.p_forenames
However this doesn't work because it includes rows that would have the n/a value in the report. A group is defined by a course code that ends in G_, however some course's don't have groups so these appear as 'n/a'.
Sorry if none of this makes sense, I just really need some help with it.
Thanks

Answer 1

I haven't entirely understood your query, but seems like you could do this in the query with a Common Table Expression (CTE):
;
WITH
CourseIsGroupCourse
AS
(
SELECT
m.m_id ,
CASE WHEN m.m_reference NOT LIKE '%G_'
THEN m.m_reference
ELSE LEFT(m.m_reference,
CHARINDEX('G', m.m_reference) - 1)
END AS course ,
SUM(...[Something I haven't figured out yet...]) as IsGroupCourse
FROM
unitesnapshot.dbo.capd_module m
GROUP BY
m.m_id,
CASE WHEN m.m_reference NOT LIKE '%G_'
THEN m.m_reference
ELSE LEFT(m.m_reference,
CHARINDEX('G', m.m_reference) - 1)
END
)
[Now back to your big query...]
SELECT
sub.course ,
sub.crs_group ,
m.m_reference ,
...
FROM
msql.unitesnapshot.dbo.capd_moduleenrolment AS me
INNER JOIN msql.unitesnapshot.dbo.capd_module AS m
ON me.e_module = m.m_id
LEFT JOIN
...
WHERE
me.e_status = 'A'
AND LEFT(m.m_reference, 2) = '12'
AND SUBSTRING(m.m_reference, 7, 2) IN ( 'VF', 'AB', 'FB' )
-- Here's the key bit...
AND sub.course in (SELECT course FROM CourseIsGroupCourse where IsGroupCourse = 1)
(When struggling with a SQL problem, I find it often helps to name tables with more descriptive names, then you can start to see what you've got where, and piece things together appropriately.)