Characteristic Relationship Challenges in IP

[Former Member]

All,

I'm having a bit of a problem with characteristic relationships in IP.

We have a simple planning cube with no data. It has the characteristics 0CALMONTH, 0MATERIAL, 0SALESORG, 0SALES_GRP, 0SALES_OFFICE, and 0PAYER. We want to build a query which allows us to plan on specifically designated valid combinations of 0SALESORG, 0SALES_GRP, 0SALES_OFFICE, and 0PAYER. However, instead of displaying only the valid combinations we desire, things seem to be showing up as the Cartesian Product of the characteristics.

We thought we could accomplish this through a characteristic relationship and we tried establishing one (first by creating a DSO and then by creating a hierarchy). But all combinations seem to yield only one of the following results:

*Only the first level of the hierarchy

*The Cartesian product

*The 2nd level of the hierarchy (and only the 2nd level) but unfortunately containing ALL options in master data

Has anyone here ever managed to do in IP what I'm trying to do (display only certain valid combinations of characteristics)? If so can you give me a step by step outline of how you did it?

According to SAP hosting we have all of the latest Service Packs.

[Former Member]

All, the issue is partially corrected. Instead of fighting with hierarchies, I stepped back and re-attacked this problem by creating a characteristic relationship based on a DSO. (I had tried this initially but it failed for a reason I will explain below.)

We do top-down planning and so I need these valid combinations at each level of granularity. I created a DSO with the key 0SALESORG, 0SALES_GRP, 0SALES_OFFICE, 0PAYER, and populated it with the valid combinations. However, when I tried to create the characteristic relationship it does not allow me to select 0PAYER as a "Source Characteristic". This results the relationship being inconsistent.

So then I created a DSO with the key 0SALESORG, 0SALES_GRP, 0SALES_OFFICE, and populated the valid combinations. It worked perfectly for my query at that level of granularity but not for the higher level of simply 0SALESORG and 0SALES_GRP and so I found myself creating another DSO and characteristic relationship to handle that level.

Now I have the valid combinations down to a certain level of granularity displaying properly, but not all the way to the 0PAYER level. I'm not sure why it has a problem aside from perhaps the fact that 0PAYER is a reference char based on 0CUSTOMER. Does anyone know why it has this problem with 0PAYER?

Additionally I have found that I can no longer execute a simple copy function on my cube because it "cannot find reference data for characteristic combination" in an insistance when I planned on a characteristic with a blank value.

[Former Member]

Randall,

As you must have seen my reply my Char Rel using Hierarchy is working fine. I am curious to know what happens when you create hierarchy as follows :

0SALESORG, 0SALES_GRP, 0SALES_OFFICE, 0PAYER.

Where 0PAYER is Source and all other elements as Target. And a Hierarchy maintained on it.

What message are you getting.

I hope you are maintaining the Hierarchy for 0PAYER and set all the other characteristics as External Chars in it.

Somehow I feel you should be able to achieve this with a Hierarchy and not DSO as mine is working...

Let me know the exact message/error you are getting with a hierarachy, so I can try it at my end..

Thanks,

Jay

[Former Member]

Jay,

I actually gave up trying to pursue the hierarchy route as I had found that creating the DSO got me closer to my goal.

The problem I was encountering when I tried to create the char relationship on the hierarchy was that the system would just create a cartesian product anywhere a '#' was included for one of the characteristics in the relationship.

With the DSO I'm able to eliminate that problem AND include valid combinations with the '#' value in certain characteristics. I still have some problems with it. I'm not able to include 0PAYER in my characteristic relationship for some reason. I also have found a scenario where I can manually plan data but when I execute a copy planning function IP throws an error ( detailed in this thread - ).

[Former Member]

Hi Randall,

I think you should give it a try using Hierarchy one more time. I will try to replicate your problem at my end on the weekend and let you know what happens on monday....

I am sure I will be able to make it work...but I need the details of your info objects...please send it to jaymankal@hotmail.com. Send me screen shots.

Take care.

Jason

[Former Member]

Gregor, I have included all of the characteristics used in my hierarchy and ONLY the characteristics used in my hierarchy but it is still giving the cartesian product.

Have you done this before in IP?

[Former Member]

Hi Randall,

Have you read the SAP Note 1091227. Interestingly it is relating your problem, but on a Multi-Provider.

Hope this helps solve your problem too..

Let us know..

J

[gregor_dieckmann]

Hi Randall,

I think that you also get automatically valid combinations since the intitial value is contained in the filter of the query, cf.

http://help.sap.com/saphelp_nw70/helpdata/en/43/1c3d0f31b70701e10000000a422035/frameset.htm

Exclude the initial values from the selection, then you will not get the automatically valid combinations.

Cf also the attached document in note 994853: Section 2.2 contains some background information about this topic.

Regards,

Gregor

[Former Member]

Mayank, in the properties I did specificy "characteristic relationships" for the aforementioned characteristics.

Jason, I do have restrictions (via a selection option variable) on each characteristic. But I still get the cartesian product. If we were only dealing with 2 characteristics this wouldn't be so bad, but with 5 it creates a ridiculous # of rows.

My basic hierarchy only has 3 Sales Orgs, 14 Sales Offices, and 34 Sales Groups. But instead of displaying 34 lines items I get 1428 lines (3x14x34).

I'm at a point where I just want to know if anyone else has actually ever done what I'm trying to do in IP.

[gregor_dieckmann]

Hi Randall,

a step in a list of characteristic relationships will only be called if all characteristics of this step are requested in the query. If not, the system will create the cart. product.

Example: Step 01 has characteristics C1,C2, C3, uses hierarchy on C3, hierarchy has the level as follows:

Level 1 C1

Level 2 C2

Level 3 C3

In the query only characteristics C1, C2 are used. Then Step 01 will not be called, the result will be the cart. product of C1, C2 values.

In a query with characteristics C1, C2, C3 (e.g. drilled down in the rows) the relation of Step 01 will be called.

Having more then one step the system computes the result per step and combines these results by an inner-join.

Regards,

Gregor

[Former Member]

Randall,

Please read online help under "Form Based Planning" in BI IP.

It states :

"Note that if you choose the Characteristic Relationships or Master Data option, a considerable number of combinations can easily be generated. For this reason, the filter restrictions for these characteristics should be accordingly restrictive. You can specify a maximum number of combinations for characteristic relationships for each real-time InfoCube."

So, you should restrict the selections again in Query Filter for each of your object- 0SALESORG, etc.

Hope this helps. If so, please post your solution so it helps others.

Jason

[Former Member]

Hi,

One small thing to be checked. In the input query, in the properties of these characteristics, have you specified 'characteristic relationships' rather than 'master data'?