Non-Cumulatives
Use
A non-cumulative is a non-aggregating key figure on the level of one or more objects, which is always displayed in relation to time. Examples of non-cumulatives include headcount, account balance and material inventory.
How you model the folder for non-cumulatives in the BI system depends on your scenario. Depending on how often non-cumulatives change, and the total number of objects for which you want to calculate non-cumulatives, you should choose one of the following two options:
Non-Cumulative Management with Non-Cumulative Key Figures:
If the non-cumulatives change infrequently, you should choose non-cumulative management with non-cumulative key figures.
If you use non-cumulative key figures, an absolute non-cumulative value (the marker) and all non-cumulative value changes are saved in the fact table of the InfoCube. In this way, the retention and volume of data in the data loading process is optimized. A data record is then only loaded into the InfoCube if a non-cumulative changes because of a transaction. Non-cumulatives can then be evaluated at any time in queries, using non-cumulative key figures.
The fact table for the InfoCube with non-cumulative key figures looks something like this (simplified):
Period |
Material |
Delta |
2003001 |
AAA |
1400 |
2003001 |
AAA |
100 |
2003002 |
AAA |
-150 |
2003003 |
AAA |
-50 |
2003004 |
AAA |
400 |
2003006 |
AAA |
-300 |
2003009 |
AAA |
1400 |
The fact table contains the transaction data. The first record in the table is the initialization. This entry does not remain physically in the table, but it is available for rebuilding the non-cumulatives.
The last record in this case is the marker (the InfoCube was compressed after this request). The non-cumulative for the period 2003005 is calculated as follows: 1400 – (-300) = 1700. This calculation takes place during the runtime of the query. Non-cumulatives can be displayed with this for time periods for which transaction data has not been loaded.
The use of non-cumulative key figures is recommended when the amount of transaction data is 20 % less than the granularity of the InfoCube.
Advantages of this Solution:
● The fact table is kept smaller.
● The history remains.
Disadvantages of this Solution:
● More administrative effort is required (for example, the InfoCube has to be compressed more often to keep the marker current).
● The query runtime can be affected by the calculation of the non-cumulative.
● Deletion of transaction data for material that is no longer current is not possible because deletion cannot be restricted by time.
See also Modeling of Non-Cumulatives with Non-Cumulative Key Figures.
Non-Cumulative Management with Normal Key Figures (Cumulative Key Figures):
If the non-cumulatives change frequently, you should choose non-cumulative management with normal key figures. That is, choose cumulative values. Absolute non-cumulatives are then retained in InfoCubes for all objects for particular key dates (for example, the end of the month). These absolute non-cumulatives can be determined from a DataStore object that is provided with non-cumulative value changes on an ongoing basis.
In this case, non-cumulative calculation takes place at query runtime. The marker is refreshed as a result of compression within the administration of an InfoCube with non-cumulative key figures.
The fact table for the InfoCube with normal key figures looks something like this (simplified):
Period |
Material |
(Delta) |
Non-Cumulative Value |
2003001 |
AAA |
(100) |
1500 |
2003002 |
AAA |
(-150) |
1350 |
2003003 |
AAA |
(-50) |
1300 |
2003004 |
AAA |
(400) |
1700 |
2003005 |
AAA |
(-) |
1700 |
2003006 |
AAA |
(-300) |
1400 |
The fact table contains the non-cumulative values, but no delta (this is here only to make it easier to understand). The non-cumulative value for a specific period can be determined using a key figure with an exception aggregation over the period.
The values for the key figures are saved in the granularity of the InfoCube. If the amount of transaction data is almost the same as the number of the most granular InfoObjects (for example, week multiplied by material) then using normal key figures is recommended.
Advantages of this Solution:
● The query runtime is not affected by the calculation of the non-cumulative.
● The deletion of transaction data for material that is no longer current is possible and easy.
Disadvantages of this Solution:
● The fact table becomes larger than it is when using non-cumulative key figures.
● In order to properly update postings in the future and the past, the data first has to be loaded into a DataStore object and then into the InfoCube.
"How to..." Paper for Inventory Management Scenarios in BI
SAP provides a document that deals extensively with the technical features of non-cumulative management in BI.
To display this document in SAP Developer Network (SDN), under https://www.sdn.sap.com/irj/sdn/howtoguides choose SAP NetWeaver 2004 → Business Intelligence → How to Handle Inventory Management Scenarios.