SQL Server AVG and Excel AVERAGE producing different results? - sql-server

Question

I'm trying to show averages on SQL server, but when I test the data in Excel the results are not the same, there must be something obvious I am missing.
Here is the code and results from SQL server:
SELECT DISTINCT
d.d_reference + ' - ' + d.d_name AS Faculty,
AVG(sub.GroupSize) AS FacultyAverage
FROM
unitesnapshot.dbo.capd_register r
INNER JOIN unitesnapshot.dbo.capd_studentregister sr ON sr.sr_register = r.r_id
INNER JOIN unitesnapshot.dbo.capd_activity a ON a.a_register = r.r_id
INNER JOIN unitesnapshot.dbo.capd_moduleactivity ma ON ma.ma_activity = a.a_id
INNER JOIN unitesnapshot.dbo.capd_module m ON m.m_id = ma.ma_activitymodule
INNER JOIN unitesnapshot.dbo.capd_department d ON d.d_id = m.m_moduledept
INNER JOIN unitesnapshot.dbo.capd_section sec ON sec.s_id = m.m_modulesection
INNER JOIN (SELECT
r.r_reference,
COUNT(DISTINCT s.s_studentreference) AS GroupSize
FROM
unitesnapshot.dbo.capd_student s
INNER JOIN unitesnapshot.dbo.capd_person p ON p.p_id = s.s_id
INNER JOIN unitesnapshot.dbo.capd_studentregister sr ON sr.sr_student = p.p_id
INNER JOIN unitesnapshot.dbo.capd_register r ON r.r_id = sr.sr_register
GROUP BY
r.r_reference) sub ON sub.r_reference = r.r_reference
WHERE
SUBSTRING(r.r_reference,4,2) = '12' AND
d.d_reference = '730'
GROUP BY
d.d_reference,
d.d_name
Here is the results in Excel:
Thanks

Answer 1

Try this for fun:
select avg(a)
from
(values(1),(2),(3),(4)) x(a);
avg(a)
-------
2
AVG() returns the same datatype as the base column. If your columns are of type int, then the result will be truncated to an int as well. The below returns the "correct" result.
select avg(cast(a as decimal(10,5)))
from
(values(1),(2),(3),(4)) x(a);
result
--------
2.5
The discrepancy you are showing (24 vs 19.50484) will most likely involve another error in conjunction with this. For example, to check that you are summing up the same data in Excel as in SQL Server, dump this result into Excel and sum it up. If it doesn't match what you currently believe is the Excel equivalent of the SQL Server data, line the columns up and check they have the same number of rows. Then sort each column individually by value ASCENDING and compare again.
SELECT d.d_name, sub.GroupSize AS FacultyAverage
FROM unitesnapshot.dbo.capd_register r
INNER JOIN unitesnapshot.dbo.capd_studentregister sr ON sr.sr_register = r.r_id
INNER JOIN unitesnapshot.dbo.capd_activity a ON a.a_register = r.r_id
INNER JOIN unitesnapshot.dbo.capd_moduleactivity ma ON ma.ma_activity = a.a_id
INNER JOIN unitesnapshot.dbo.capd_module m ON m.m_id = ma.ma_activitymodule
INNER JOIN unitesnapshot.dbo.capd_department d ON d.d_id = m.m_moduledept
INNER JOIN unitesnapshot.dbo.capd_section sec ON sec.s_id = m.m_modulesection
INNER JOIN (SELECT r.r_reference,
COUNT(DISTINCT s.s_studentreference) AS GroupSize
FROM unitesnapshot.dbo.capd_student s
INNER JOIN unitesnapshot.dbo.capd_person p ON p.p_id = s.s_id
INNER JOIN unitesnapshot.dbo.capd_studentregister sr ON sr.sr_student = p.p_id
INNER JOIN unitesnapshot.dbo.capd_register r ON r.r_id = sr.sr_register
GROUP BY r.r_reference) sub ON sub.r_reference = r.r_reference
WHERE SUBSTRING(r.r_reference,4,2) = '12' AND d.d_reference = '730'
ORDER BY d.d_name