How to calculate query value on extended range of data

Question

Hi,
This is my first post, so let me greet all forum users. I'm reading the forum for few weeks and I'm really impressed with the force of this community.
My question is based on some real problems but I'll ask in general:
In a BEx query - is it possible (and how?) in a result cell to obtain value that is calculated based on some more extended range of infoprovider records than it results from all the characteristics related to that cell?
The question is somehow about a reverse case to the restriction in Selection. I'm asking about a kind of 'extension' of selection.
Example for the question follows (it is somehow simple, I put it here just to picture the problem, note: the question is more general!)
cube: ch: 0CALMONTH, 0MATERIAL, 0PLANT
      kf: 0QUANTITY (ex. of outgoing deliveries)
What I need is to obtain in any query the SUM(0QUANTITY) over all 0PLANT (this can lead to knowing the 'activity' of the specific plant by calculating the share. that activity can be calculated at diferent level of detail).
The 'dream' query would look like:
rows: 0CALMONTH, 0MATERIAL, 0PLANT      
cols: 0QUANTITY  SUM_OVER_0PLANT
resulting in ex:
0CALMONTH  0MATERIAL  0PLANT  0QUANTITY  SUM_OVER_0PLANT  PLANT_ACTIVITY
2005.01    00001      P100    10         30               1/3
2005.01    00001      P200    20         30               2/3
2005.01    00002      P100    30         70               3/7
2005.01    00002      P200    40         70               4/7
2005.02    00001      P100    50         110              5/11
2005.02    00001      P200    60         110              6/11
2005.02    00002      P100    70         150              7/15
2005.02    00002      P200    80         150              8/15
after removing drill 0MATERIAL:
0CALMONTH  0PLANT  0QUANTITY  SUM_OVER_0PLANT  PLANT_ACTIVITY
2005.01    P100    40         100              4/10
2005.01    P200    60         100              6/10
2005.02    P100    120        260              12/26
2005.02    P200    140        260              14/26
after removing drill 0CALMONTH:
0PLANT  0QUANTITY  SUM_OVER_0PLANT  PLANT_ACTIVITY
P100    160        360              16/36
P200    200        360              20/36
For this specific case I investigated several solutions:
1) using formula & function SUMCT (Result). This has a drawback that Result for 0PLANT must be present in right place (bottom level) in the resulting query. I don't like this.
2) SUM can be directly precalculated in the cube. Well, I'd prefer Bex only...
Please keep in mind also following:
1) there may be a need to refer to the data that is outside prompt/filtering/restricting range.
2) I want to preserve all the OLAP freedom, so all the solution should be a query with some tricky formula/kf/??? to be used by user in any situation and producing right result.
I hope I'm not demanding to much...
(now after this long example please have a look at the question again!)
Regards,
Mirek

Answer 1

Hi Ashwin,
while using SUMCT I loose the freedom of OLAP. SUMCT calculates properly SUM_OVER_0PLANT only when the 0PLANT is the lowest drillin level. If this is not the case (ie. some other ch is the lowest level or 0PLANT results are suppressed) the approach won't provide right result.
sure I will reward all helpfull posts.
regards,
Mirek