Quantcast
Channel: SCN : All Content - SAP Business Warehouse
Viewing all articles
Browse latest Browse all 5981

Time Distribution: Where art thou in BW v7.0?

$
0
0

Hello All,

 

I have a requirement where I feel Time Distribution would be suitable to deliver a solution. But where is this functionality in v7.0?

 

I have cube where some data is at day level (i.e. 0CALDAY is populated) and some data is not at day level i.e. 0CALDAY is not populated. 0CALWEEK, 0CALMONTH, 0CALYEAR are populated.

 

This scenario has raised a challenge: I need the data from this cube however the data I pull must be against the Gregorian calendar. (e.g. When I ask for March 2014's data It must be transaction data between 1st March thru to 31st March).

 

For those records in the cube that are at day level that isnt a problem: a select against 0CALDAY from 01.03.2014 to 31.03.2014 will suffice.

But as I noted earlier, 0CALDAY isnt always populated. So using 0CALDAY to select a months worth of data isnt ideal.

 

Sadly 0CALMONTH cannot be used because when assigning the value for 0CALMONTH, (and 0CALWEEK, 0CALYEAR) the 4-4-5 calendar is used NOT the Gregorian Calendar. This means that if 0CALMONTH is 03.2014 this covers data from 2nd thru to 29th March. 1st March data will be under 02.2014. Transaction data for 30th and 31st March fall under 04.2014. So a select on the cube where 0CALMONTH = 03.2014 wont meet the requirement.

 

My belief is that I should consider the 'Time Distribution' concept:

Move data from the original cube to a new ODSO (ZO_B) which is at 0CALMONTH level. Now that data is aggregated at month level, move it to a new cube ZC_B and populate 0CALWEEK whereby the key-figure ZACTSLSVAL is the the distribution of the month value:

If ZACTSLSVAL in the ODSO is 1000 EUR for March 2014 then for each of March's weeks the Key Figure value is as follows:

09 =  (2/31)x1000 = 64.52

10 = (7/31)x1000 = 225.81

11 = (7/31)x1000 = 225.81

12 = (7/31)x1000 = 225.81

13 = (7/31)x1000 = 225.81

14 = (1/31)x1000 = 32.26

 

I can now extract from cube ZC_B, ask for March 2014 data and be confident that the key figure returned is for the value as per the Gregorian Calendar.

 

Unfortunately, it seems that the 'Time Distribution' mapping in v3.5 Update Rules isnt available in v7.0 Transformations!

I havent found a definitive answer in terms of where this functionality is in v7.0.

 

Can you offer advice in terms of where this functionality is and also regarding the requirement described above and my approach.

Many thanks,


Viewing all articles
Browse latest Browse all 5981

Trending Articles



<script src="https://jsc.adskeeper.com/r/s/rssing.com.1596347.js" async> </script>