I’m developping a report with multiple tables and graphs where the section is used as a bookmark for reportbursting into several pdf files.
So far the last obstacle is the running average. This works perfectly, just not when there is a section added.
In my example in EFashion, i have used following variables and a section on State.
running avg year N : =NoFilter(Previous(Previous(RunningSum([Sales revenue]);0)-Previous(RunningSum([Sales revenue]);12);0)/12)
running avg year N-1 : =NoFilter(Previous(Previous(RunningSum([Sales revenue]);0)-Previous(RunningSum([Sales revenue]);12);12)/12)
i used sum N and sum N-1 to add a context to the variables
sum 2012 : =(Sum([J(N) 12 month runavg] In([State];[Year];[Month])))
sum 2011 : =(Sum([J(N-1) 12 month runavg] In([State];[Year];[Month])))
but somehting goes wrong in the context I suppose. (might not be the shortest way of defining variables btw)
Webi now calculates the running average over the LAST 12 LINES.
for N-1 → calculates over partly lines CURRENT SECTION AND PARTLY OVER PREVIOUS SECTION …
→ see State ‘Colorado’ : the average is far to high, as webi calculates over revenues of ‘California’ as well.
how can i fix this? how do i define the context correctly so webi only calculates within the current section?
file should be added (new here, so please reply if i didn’t do so correctly) RunAVG EFashion.wid (107.0 KB)
i found that there is a Reset_Dimenstion needed.
for a simple runningsum this would be
=RunningSum([Sales revenue];([State]))
this works fine, but in the running average formulas this doesn’t work.
i’ve tried to imply this as followed : (for a 3 month average this time, is more apprehensive than 12 month)
However, when “NoFilter” is not added to the formula and 2003 is selected for example, the average is of course calculated on the filtered data… not what im looking for.
is there anyway to combine those? or is that simply not possible in BO?
not really no.
i could contact the firm that maintains the DWH but i suppose they are not so keen on looking into custom variables and formulas like these.
could there be a different approach than the nested previous functions that works better?
anyway, cant be i’m the first person trying this right?