I've a problem with cumulative values in two scenarios: Stock changes and Turnover, both by month.
A wide varity of articles results in not having fact data for all articles and all months.
Basically, using cumulative feature of QV Chart seems to be the solution, but it shows wrong overall month totals: the cumulated values of articles that have no value in a month are missing in that monts.
As a simplified example it looks like this, using "full acumulation" and "sum(value)" or "value" as formula:
As there's no value for b (red) in November, b is not mentioned at this dimension element. The cumulated total for 11.11.2011 is simply wrong; it should be 400!
One can apply a dirty fix by adding values with less significant decimals for any article and any month, but as there are multiple dimensions in the real app and the non-moving articles are the majority by far, this would let the data volume explode.
Does anyone know a best practise for that? QV should be able to show right figures when it offers a cumulation feature out-of-the-box.
The QVW for this excample is attached and a second row of charts shows the dirty-fixed example.