case statement trying to aggregate data in case statement using aggregate functions - sql

Question

SELECT c.DESCRIPTION,
fobt.OBJECT_ID,
fobt.FA_OBJECT_TRANS_VALUE_TYPE,
fobt.OBJECT_TRANSACTION_ID,
fobt.OBJECT_TRANSACTION_DATE,
fobt.AMOUNT,
fobt.FA_YEAR,
fobt.FA_PERIOD,
SUM(fobt.AMOUNT) OVER(
PARTITION BY fobt.
ORDER BY fobt., fobt.OBJECT_TRANSACTION_DATE
) AS RUNTOTAL,
DENSE_RANK() OVER(
PARTITION BY fobt.OBJECT_ID
ORDER BY fobt.FA_YEAR DESC
) RK1,
DENSE_RANK()
OVER(
PARTITION BY fobt.OBJECT_ID, c.DESCRIPTION
ORDER BY
fobt.FA_YEAR,
fobt.FA_PERIOD,
c.DESCRIPTION DESC
)
rnk2
FROM FA_OBJECT_TRANSACTION fobt
INNER JOIN (SELECT DISTINCT COMPANY,
OBJECT_ID,
OBJECT_TRANSACTION_ID,
CODE_C
FROM FA_) FT
ON fobt. = FT.
AND fobt. = FT.OBJECT_TRANSACTION_ID
INNER JOIN CODE_C c
ON FT.CODE_C = c.CODE_C
AND FT.COMPANY = c.COMPANY
WHERE fobt.OBJECT_ID = 12345
AND fobt.FA_ IN ('Depreciation')
ORDER BY 2, 4
I'm trying to fix the DENSE_RANK for OBJECT_ID, DESCRIPTION, YEAR AND PERIOD. There is only one object id = 1234 but there are 2 locations (Rotterdam and South Windsor). Equipment A was logged in location South Windsor from 5/2015 to 3/2016 and then transferred to Rotterdam on 4/2016. I will like to have a DENSE_RANK by PARTITION by OBJECT_ID and LOCATION but the order by is not working properly. I would like to take the latest date(year and month by location of that piece of equipment. For example, Equipment A on 3/2016 at location South Windsor would be ranked 1 because it is the latest date at that location and then transferred to location Rotterdam on 4/2016 and that would also get tagged ranked 1 because it changed location. Below are the current results and second example is the desired result.

Answer 1

If I understood everything correctly You need to add only DESC after fa_year in last dense_rank:
... dense_rank() over(partition by object_id, description
order by fa_year desc, fa_period desc) as rk2
Alternatively use over (... order by transaction_id desc).
Test:
with t (DESCRIPTION, OBJECT_ID, transaction_id, AMOUNT, FA_YEAR, FA_PERIOD)
as (select 'South Windsor, CT', 22208, 278620, 257, 2015, 5 from dual union all
select 'South Windsor, CT', 22208, 287864, 257, 2015, 6 from dual union all
select 'South Windsor, CT', 22208, 305508, 257, 2015, 7 from dual union all
select 'South Windsor, CT', 22208, 322323, 257, 2015, 8 from dual union all
select 'South Windsor, CT', 22208, 330902, 257, 2015, 9 from dual union all
select 'South Windsor, CT', 22208, 340183, 257, 2015, 10 from dual union all
select 'South Windsor, CT', 22208, 350218, 257, 2015, 11 from dual union all
select 'South Windsor, CT', 22208, 359261, 257, 2015, 12 from dual union all
select 'South Windsor, CT', 22208, 374836, 257, 2016, 1 from dual union all
select 'South Windsor, CT', 22208, 382864, 257, 2016, 2 from dual union all
select 'South Windsor, CT', 22208, 388910, 257, 2016, 3 from dual union all
select 'Rotterdam, NY', 22208, 388915, 257, 2016, 4 from dual)
select description, transaction_id, fa_year, fa_period,
sum(amount) over (partition by object_id order by fa_year, fa_period) total,
dense_rank() over(partition by object_id order by fa_year desc) rk1,
dense_rank() over(partition by object_id, description
order by fa_year desc, fa_period desc) rk2
from t
order by object_id, transaction_id
Output:
DESCRIPTION TRANSACTION_ID FA_YEAR FA_PERIOD TOTAL RK1 RK2
----------------- -------------- ---------- ---------- ---------- ------ ------
South Windsor, CT 278620 2015 5 257 2 11
South Windsor, CT 287864 2015 6 514 2 10
South Windsor, CT 305508 2015 7 771 2 9
South Windsor, CT 322323 2015 8 1028 2 8
South Windsor, CT 330902 2015 9 1285 2 7
South Windsor, CT 340183 2015 10 1542 2 6
South Windsor, CT 350218 2015 11 1799 2 5
South Windsor, CT 359261 2015 12 2056 2 4
South Windsor, CT 374836 2016 1 2313 1 3
South Windsor, CT 382864 2016 2 2570 1 2
South Windsor, CT 388910 2016 3 2827 1 1
Rotterdam, NY 388915 2016 4 3084 1 1