BW Query Consolidation

[Former Member]

Hi Experts,
  I have a typical scenario where we are in need of 
Query level consolidation with the existing cube 
design.The scenario is

Type  Assigned  Owner  KF1 KF2 KF3 KF4

A1      Bob             1   2
A2               Bob            3   4
A1      Rob             5   6
A2               Rob            7   8

I want it to be consolidate with the characteristics 
which can be a new structure and if assigned and owner 
are same,

Assigned/owner   KF1  KF2  KF3  KF4
Bob               1    2    3    4
Rob               5    6    7    8

How can this be handled?Any documents will be appreciated

Thank you
PA

[Former Member]

Hi,

We have the existing cube with the above specified values and i have added an new characteristics assigner/owner in Cube and the ODS(Enterprise Data warehouse). I have made my routine to fill the assigner/owner in the update rule of the ODS so that it get overwritten.

I just want to understand, will this new field gets populated only for the next loaded delta or will it also change the existing history data?

What happens to the data in Infocube.Will this Assigner/owner field gets updated for the old existing data?Usually insert is executed at the infocube level.

Thank you

arun

[Former Member]

Arun,

infocube is not a transparent table as on ODS, so no overwrite is possible (append and then OLAP aggregation)....let's say that your delta will come up with your new field filled, but all the history remains as in the past (in your query you will see some records with assigner/owner not assigned and something not!)

Bye,

Roberto

[Former Member]

How about this..Create a new virtual cube with all characteristics and keyfigures of OLD cube. Connect these two by update rules...in the update rules derive one char based on other two...

With this solution: Your existing cube is untouched,

you have to create new virtual cube,,

you have export datasource of oldcube,

you have created update rules,

you have create a query on top of virtual cube..

Anil

[Former Member]

Hi,

Thanks for your reply. I don't understant how to use Virtual Characteristics. Will it be part of the Infocube data model.Can someone explain with this scenario?

How will the data get filled from the existing infocube data? Should the data load need to be rescheduled?

Can you forward me to some link regarding the topic 'Virtual characteristics' ?

Thank you

PA

[Former Member]

Hi,

you can find a lot of info if you search in these forums with 'virtual' as key word and more in the system documentation; choose the sub-object 'Documentation' in transaction SMOD for enhancement RSR00002 (BW: Virtual Characteristics and Key Figures in Reporting) and then choose 'Display'.

Hope it helps !

Bye,

Roberto

[Former Member]

you will have to add a characteristic in the cube but you don't have to reload the cube. This new characteristic is empty in the cube, has to be in the cube to be able to select it in the query, but is filled by a user exit when running the query.

[Former Member]

the easiest and most efficient (from query performance point of view) is to add a new characteristic in your cube.

With the "existing cube design", the only way I can think of is to create a virtual characteristic (you can find documents, example on sapnet in the how to guides section, it's a user exit).

[GFV]

Hi,

if I understand your problem you would like to see in the same column Assigned and owner just if they are the same person ... and 4 KFs: 1 and 2 are delimited by Type A1 while 3 and 4 by Tpe A2.

The problem is making a test on two different Chars Assigned and Owner ... According to me the best solution is a modification on data flow (eg a tech Char that is true if Assigned = Owner and false if not ...).

Obviously you can work even with Virtual characteristics, but with no good performance, or VBA.

Hope it helps

GFV