SSRS - calculate percent change with variable years and number of years - reporting-services

Question

Hopefully, I can explain this clearly! Our report allows user to select a variable number of academic years. They can be consecutive or not. In this case, say the user selects 2017, 2015, 2013. The report presents like this:
AcYear TotCredits %Change
2013 251 0.00
2015 255 0.00
2017 1102 0.00
Because the SQL that provides the datasource is assuming the previous year is one year prior. So 2015 is being compared against 2014, even though we're not selecting that info, which is why it's 0. 2017 is being compared against 2016, even though the user selected 2015. So, either in the initial SQL or in the report table expression, how would I go about getting the credits for the most immediately prior SELECTED academic year, no matter how many were selected, and then calculating the percent change based on that??
(If the user selected consecutive years, say 2017 and 2016, the data looks like this:)
AcYear TotCredits %Change
2016 458 0.00%
2017 19 -95.85%
This is the self join to get the "previous" year:
from
cte
left join cte as prev1
on cte.academic_year - 1 = prev1.academic_year
and cte.subject = prev1.subject
Thanks for any ideas!

Answer 1

You could try to use previous function to see whether it works or not. You need to make sure Year has been sorted. Below is design
select * from (select 2012 as year, 12 as amount
union all
select 2013 as year, 5 as amount
union all
select 2014 as year, 6 as amount
union all
select 2015 as year, 4 as amount
union all
select 2016 as year, 24 as amount)tt
where year in (#t)
Expression:
=iif(Fields!year.Value-previous(Fields!year.Value)=1,(Fields!amount.Value-previous(Fields!amount.Value))/iif(IsNothing(previous(Fields!amount.Value) ), 0,previous(Fields!amount.Value)),0)
Then you will get the result like below
Zoe

Answer 2

Surprisingly easy/elegant fix, just took me a long time to find it. The original import bit:
SELECT
cte.academic_year
, cte.subject
, cte.subject_desc
, cte.credits as cur_credits
, prev.credits as prev_credits
FROM
cte
LEFT JOIN cte as prev
on cte.academic_year - 1 = prev.academic_year
and cte.subject = prev.subject
The new improved code:
SELECT
t.academic_year
, t.subject
, t.subject_desc
, t.credits as cur_credits
, prev_credits = prev.credits
FROM
cte t
OUTER APPLY
(SELECT top 1 credits from
cte
WHERE t.academic_year > academic_year
and t.subject = subject
ORDER BY academic_year desc) prev
It would have been very nice to use LAG, but alas, we are on SQL2008.