Multiple Aggregates Pivot Table - sql

Question

I have a few columns of data I am wanting to put into a Pivot table. This is what the data currently looks like once the following query is ran:
SELECT
t.clinic,
t.fiscal_year,
t.total,
n.new_pats,
(t.total - n.new_pats) AS active
FROM (SELECT DISTINCT
clinic,
fiscal_year,
COUNT(DISTINCT patient_id) AS total
FROM transactions t
JOIN period p
ON (t.date_entered BETWEEN p.period_start AND p.period_end)
GROUP BY clinic, fiscal_year) t
JOIN (SELECT DISTINCT
clinic,
per.fiscal_year,
COUNT(DISTINCT patient_id) AS new_pats
FROM patient pat
JOIN period per
ON (pat.first_visit_date BETWEEN per.period_start AND per.period_end)
GROUP BY clinic, fiscal_year) n
ON (t.clinic = n.clinic AND t.fiscal_year = n.fiscal_year)
And I would like the data to be broken up by the three aggregates in their own rows and the years being the columns where it looks something like this:
This could also be broken into 3 separate columns per aggregate per year, but this is what I'm ideally aiming for. I haven't put together a Pivot table within SQL before and am at a loss. Is there a better way to go about formatting this data in the desired way?

Answer 1

You could try this
SELECT clinic,
aggregate_field,
ISNULL([2009],0) AS [2009],
ISNULL([2010],0) AS [2010],
ISNULL([2016],0) AS [2016],
ISNULL([2017],0) AS [2017]
FROM (SELECT clinic,
fiscal_year,
aggregate_field,
value
FROM (
------- This part is your original query -----------------
SELECT t.clinic,
t.fiscal_year,
t.total,
n.new_pats,
( t.total - n.new_pats ) AS active
FROM (SELECT DISTINCT clinic,
fiscal_year,
Count(DISTINCT patient_id) AS total
FROM transactions t
JOIN period p
ON ( t.date_entered BETWEEN
p.period_start AND p.period_end )
GROUP BY clinic, fiscal_year) t
JOIN (SELECT DISTINCT clinic,
per.fiscal_year,
Count(DISTINCT patient_id) AS new_pats
FROM patient pat
JOIN period per
ON ( pat.first_visit_date BETWEEN
per.period_start AND per.period_end )
GROUP BY clinic, fiscal_year) n
ON ( t.clinic = n.clinic
AND t.fiscal_year = n.fiscal_year )
-------------------------------------------------------------
)
unpivot_source
UNPIVOT ( value
FOR aggregate_field IN (total,
new_pats,
active) ) unpivot_result) AS
pivot_source
PIVOT ( Max(value)
FOR fiscal_year IN ([2009],
[2010],
[2016],
[2017]) ) AS pivot_result
ORDER BY clinic, aggregate_field DESC
I've create a demo here http://rextester.com/MFHWV68715.

Answer 2

Try this:
CREATE TABLE #ActualData
(
Clinic varchar(50),
fiscal_year int,
total int,
new_pats int,
active int
)
INSERT INTO #ActualData VALUES ('A', 2016, 3538,1787,1751), ('A', 2017, 1218,373,845), ('B', 2009, 1,2,3), ('B', 2010, 1,2,3)
SELECT * FROM #ActualData ad
;WITH temps AS
(
SELECT Clinic, fiscal_year, aggregateF, Number FROM
(
SELECT ad.Clinic, ad.fiscal_year, ad.total,ad.new_pats,ad.active FROM #ActualData ad
) src
UNPIVOT
(
Number FOR aggregateF IN (total,new_pats ,active)
) unpvt
)
SELECT Clinic, aggregateF, [2009],[2010],[2016],[2017]
FROM
(
SELECT t.Clinic, t.aggregateF, t.fiscal_year, t.Number FROM temps t
) src
PIVOT
(
MIN (Number) FOR fiscal_year IN ([2009],[2010],[2016],[2017])
) pvt
ORDER BY pvt.Clinic, pvt.aggregateF DESC
DROP TABLE #ActualData