Disaggregation "doubling" up the value partial week-month

[Former Member]

Dear Experts,

A bit of weird case for which I cant find any config level solution.

KF1 is disaggregated based on KF2, both structurally and along time.

KF2 has value= 0 (not blank) across Jan14 to Dec14.

I enter a value in KF1 in TB =  months e.g. Jan14 = 500, Feb14 = 600, March14 = 900

I view the values of KF2 in TB weeks and this is how it looks on data view = weeks

W.06.01.2014	100
W.13.01.2014	100
W.20.01.2014	100
W.27.01.2014	220
W.03.02.2014	120
W.10.02.2014	120
W.17.02.2014	120
W.24.02.2014	300
W.03.03.2014	180
W.10.03.2014	180

In the weeks split across consecutive months, I note that the values are sum of values of weeks on the either side

It is obvious to see what is happening but why will this happen and how to get rid of this obvious bad math.

Thanks

BS

PS: Storage bucket profile is weeks and months.

[former_member187488 · Accepted solution]

Hi, you need to understand the concept of technical periods. Refer to note 737230.

Data in liveCache is saved in technical periods, instead of the periods you defined in storage bucket profile.

Technical periods are derived from the storage bucket profile definition.

In your case, the periods are:

As you can see, M01.2014 and M01.2015 includes 5 technical periods, so the data you input is divided evenly to each technical periods (because all proportional factors are zeros.) Thus the five periods in Jan have value 100 (500/5), while the 5 periods in Feb have value 120 (600/5).

Then for week W 27.01.2014, it includes 2 technical periods, so the value is 100 + 120 = 220.

You can calculate for March by yourself

[Former Member · Accepted solution]

Thanks Ada,

I have said the same thing above (it is obvious what is happening.... ) and I am challenging this logic of SAP:-). if you allow me I can write the pseudo code for this

Thanks and appreciate your attention.

Regards

BS

[former_member187488 · Accepted solution]

HHI, this is not bad math, this is system logic. Everything is because you are using proportional factor, but you did not set proper values for the proportions. As long as you have this setting, you should be aware of the result. If you do not want this behavior, you should make other settings. Disaggregation is done in liveCache, you cannot control it by ABAP coding.

[Former Member · Accepted solution]

Disaggregation is done in liveCache, you cannot control it by ABAP coding.

Thanks Ada,

That's the right answer. So the code of the COM routine needs to be changed else SAP should mention in the help and context documentation or a dialog pop-up that results may be wrong if the basis key figure used for disaggregation has 0 values, else there is no way to imagine this scenario except testing 500 odd case conditions. Humanly impossible to do, unless there is a separate state budget for doing this.. something like Obama care.

There is no denying that this way of calculation is not acceptable to any customer in the world. Atleast I cant think of one in my tiny lil mind 🙂

Thanks

BS

[Former Member · Accepted solution]

Ada,

Sorry I have to take back the points here.

This is indeed a BUG whether you like to admit it such or not.

When the reference key figure KF2 for dis-aggregation OF KF1 (P, K) is 0 (as in 0 typed) then the behavior is like above.

If the reference key figure is blank (not 0), then the behavior is correct or if I may say acceptable. In this case the the value of KF1 in the week spread over 2 consecutive months doesn't show the red numbers above. It is split proportionately.

This means SAP can do something about this. EVEN IF the reference key figure KF2 is 0, the split of KF1 should be same as it happens when KF2 is blank.

I cant imagine any other logic or rationale at this point in time for the behavior I complained above. My conclusion is based on testing in a "standard" environment.

Thanks

BS

[Former Member · Accepted solution]

Borat,

Did you ever log this issue with SAP and was there a resolution offered by them?  I am experiencing the same issue and it looks as though the trigger is having the "zero allowed" indicator set for the KF.  IMO you are correct that this is something that SAP needs to address.

Thanks

Sean

[Former Member · Accepted solution]

Sorry Sean,

I have let the client IT know that this is a bug and that they should file for damages with SAP if they don't come back with a program correction.

Do publish your discoveries. An even simpler question I had was for what sort of key figures (by business meaning and use) should one have zero allowed checked ?. I have stopped bothering about the answers after having found no clear answers on help.sap.com :-). There must be some intuitive way of doing this, I am sure only SAP has the answers and assumptions!

Thanks

Borat

[satish_waghmare3 · Accepted solution]

Hello BS,

I am not sure if you have seen this SAP Note#1902413  ( Disaggregation incorrect?  - How to analyze time based disaggregation and see the internal time based disaggregation factors )

Please read section where it is mentioned as "Some important points to consider"

May be this will help.

Thank you

Satish Waghmare

[Former Member · Accepted solution]

Satish,

I had actually seen this note and the /SAPAPO/OM_TS_BROWSER program showed the bucket breakdown to be as expected (i.e. that the split week was indeed split into 2 storage buckets with a different day count for each).  That said, this is a good resource for anyone looking at time disaggregation problems, so thank you for that.

Sean

[Former Member · Accepted solution]

Thank you Satish for sharing this well illustrated note...( that is kind of uncommon to have from SAP :-))

My search for 5 non stop hours some months ago, did not pop up this note. I used all kinds of key words incl. simple ones like "time based disaggregation"

Here is what SAP says for the record of everybody.

If all the values are ZERO (and NOT initial), each  bucket is treated equally (which results in an equal disaggregation among the single time buckets).

Now with my knowledge of common English, I can't possibily relate this statement to the observation above.

Secondly, this logic according to me is a flawed logic. If SAP is applying the days in a calendar month logic to apportion the values to the week (bucket) as explained in the example in the note, the logic should work the same whether the reference proportioins are ZERO or Initial (Blank). This is becuase the internal proportions are based on days in a month !! Why should it matter if reference key figure is zero or blank 🙂 May be I didnt understand this fully with my tiny lil mind. May be there are side effects not tested by SAP or anyone else.

PS: SAP should employ the services of a firm like google or yahoo for making note search less torturous. An even better idea is the notes relevant for each program, application, key field, tables and module pool programs are embebbed in the help objects (F1), so that the users can click on the links to all notes relevant to the release and SP. I wonder why SAP won't do it. Probably it is a 10 year project given it has not been done all these years. My firm can volunteer if SAP wants it 🙂 Note search is a colossal waste of time.

Thanks and keep educating.

BS

[former_member187488 · Accepted solution]

Hi Borat,

It has been some time before I have time to visit our forum again ... though I've received thousands of mails from the forum ...
The "uncommon" note is from our very experienced architect. Good to know you like that (so do I).

I could not comment more on your comment about disaggregation logic ... you can create your development request to SAP if you want ...

As far as I know, you can try to contact  our "customer experiencing team" for your great idea about note search ...

See you later here

Best Regards,
Ada

[Former Member · Accepted solution]

Thank you Ada

Nice to know about this customer experience team. I didnt know such a thing existed. How do I reach out to them directly for passing on a tip or two for free :-). Any channel, weblinks, emails, phones? I have a laundry list to complain.

I have sent this discussion link to my previous employer. They will raise the request for the mankind with a word not call it "customer specific" development :but as a program correction with medium priority for general release to all customers, valid for all releases.

Danke

BS

[former_member187488 · Accepted solution]

Hi Borat,

Sorry that I only know we have such a team, but don't know how to contact them ...
Maybe you can ask the sales people?

BR, Ada

[Former Member · Accepted solution]

Hello,

just by chance I stumbled upon this, may be you should give it a try (didn't know this existed at all)

http://influence.sap.com

regards,

J.

[Former Member · Accepted solution]

Thanks Ada, James,

I am scared of sales people. They managed to sell me things we never used. I am such an easy customer 🙂

Zuper.

Danke

BS