How do I add a moving window to rank() using Oracle SQL - sql

Question

I need to rank values over a moving time window. I got some direction from a blogpost by Tony Hasler at
https://tonyhasler.wordpress.com/2012/10/24/model-clause-use-cases/#comment-5116 but the solution of adding a windowing clause to non-windowing functions like median does not work with the rank() or percent_rank() functions which are analytic non-windowing functions.
Working example with median() function:
with a as (
select 'a' sector, trunc(sysdate) dt, 64 v from dual union all
select 'a' sector, trunc(sysdate)-1 dt, 2 from dual union all
select 'a' sector, trunc(sysdate)-2 dt, 4 from dual union all
select 'a' sector, trunc(sysdate)-3 dt, 128 from dual union all
select 'a' sector, trunc(sysdate)-4 dt, 8 from dual union all
select 'a' sector, trunc(sysdate)-5 dt, 16 from dual union all
select 'a' sector, trunc(sysdate)-6 dt, 32 from dual union all
select 'a' sector, trunc(sysdate)-7 dt, 256 from dual union all
select 'a' sector, trunc(sysdate)-8 dt, 1 v from dual union all
select 'a' sector, trunc(sysdate)-9 dt, 512 from dual union all
select 'b' sector, trunc(sysdate) dt, 3 from dual union all
select 'b' sector, trunc(sysdate)-1 dt, 27 from dual union all
select 'b' sector, trunc(sysdate)-2 dt, 9 from dual union all
select 'b' sector, trunc(sysdate)-3 dt, 81 from dual
)
select * from a
model
partition by (sector)
dimension by (dt)
measures (v, 0 mov_rank)
rules
(
mov_rank[ANY] = median(v)[dt between CV()-3 and CV()]
)
order by sector, dt
;
The example does not work if we replace median with rank() as in:
with a as (
select 'a' sector, trunc(sysdate) dt, 64 v from dual union all
select 'a' sector, trunc(sysdate)-1 dt, 2 from dual union all
select 'a' sector, trunc(sysdate)-2 dt, 4 from dual union all
select 'a' sector, trunc(sysdate)-3 dt, 128 from dual union all
select 'a' sector, trunc(sysdate)-4 dt, 8 from dual union all
select 'a' sector, trunc(sysdate)-5 dt, 16 from dual union all
select 'a' sector, trunc(sysdate)-6 dt, 32 from dual union all
select 'a' sector, trunc(sysdate)-7 dt, 256 from dual union all
select 'a' sector, trunc(sysdate)-8 dt, 1 v from dual union all
select 'a' sector, trunc(sysdate)-9 dt, 512 from dual union all
select 'b' sector, trunc(sysdate) dt, 3 from dual union all
select 'b' sector, trunc(sysdate)-1 dt, 27 from dual union all
select 'b' sector, trunc(sysdate)-2 dt, 9 from dual union all
select 'b' sector, trunc(sysdate)-3 dt, 81 from dual
)
select * from a
model
partition by (sector)
dimension by (dt)
measures (v, 0 mov_rank)
rules
(
mov_rank[ANY] = rank() over (order by v)[dt between CV()-3 and CV()]
)
order by sector, dt
;
I would appreciate any help.
Thanks.

Answer 1

This might be a little "old-fashioned", but you might be able to achieve an equivalent result using a self-join instead of analytics or model, as in something like:
with a as (
select 'a' sector, trunc(sysdate) dt, 64 v from dual union all
select 'a' sector, trunc(sysdate)-1 dt, 2 from dual union all
select 'a' sector, trunc(sysdate)-2 dt, 4 from dual union all
select 'a' sector, trunc(sysdate)-3 dt, 128 from dual union all
select 'a' sector, trunc(sysdate)-4 dt, 8 from dual union all
select 'a' sector, trunc(sysdate)-5 dt, 16 from dual union all
select 'a' sector, trunc(sysdate)-6 dt, 32 from dual union all
select 'a' sector, trunc(sysdate)-7 dt, 256 from dual union all
select 'a' sector, trunc(sysdate)-8 dt, 1 v from dual union all
select 'a' sector, trunc(sysdate)-9 dt, 512 from dual union all
select 'b' sector, trunc(sysdate) dt, 3 from dual union all
select 'b' sector, trunc(sysdate)-1 dt, 27 from dual union all
select 'b' sector, trunc(sysdate)-2 dt, 9 from dual union all
select 'b' sector, trunc(sysdate)-3 dt, 81 from dual
)
select
a.sector,
a.dt,
a.v,
count(case when self.v < a.v then self.v end) + 1 mov_rank
from
a,
a self
where
self.sector = a.sector
and
self.dt between a.dt - 3 and a.dt + 3
group by
a.sector,
a.dt,
a.v
order by
a.sector,
a.dt,
a.v;