Modeling Non-Cumulatives with Non-Cumulative Key Figures
A non-cumulative key figure is modeled in BW using the corresponding field for the non-cumulative value change or the corresponding fields for inflows or outflows. You can determine the current non-cumulative or the non-cumulative at a particular point in time. To do this, you use the current, end non-cumulative and the non-cumulative changes and/or the inflows and outflows.
The purpose of the special folder for non-cumulative key figures is to optimize data transport into the BW system, to retain data, and to access the database when evaluating in reporting in the BW system.
Whether the modeling of non-cumulatives with non-cumulative key figures is useful depends on your scenario. We recommend using non-cumulative key figures for areas where non-cumulatives do not regularly change completely, for example with warehouse stock (retail) or with the number of employees. For more information, see the detailed comparison in Non-Cumulatives.
Model of Non-Cumulative Key Figures
A non-cumulative InfoProvider - an InfoCube or DataStore object (advanced) - is modeled with at least one non-cumulative key figure. These non-cumulative key figures are mapped using one key figure for non-cumulative changes or two key figures for inflows and outflows. The option you choose depends on how you want to evaluate the non-cumulative key figure.
The key figures for non-cumulative value change or for inflows and outflows are normal cumulative key figures that have 'summation'both as aggregation and exception aggregation. Non-cumulative key figures always have summation as standard aggregation (aggregational behavior on the database, for example, upon compression or roll-up of aggregates). However, with reference to a time characteristic, they have an exception aggregation (in reporting) that is not equal to summation, as it would not make sense to cumulate non-cumulatives by time.
Non-cumulative key figures such as Number of Employees are cumulated using characteristics such as Cost Center. There is not much point in totalling the number of employees using different periods however. With periods, you can form an average for example.
With the cumulative value Sales Revenue for example, it makes sense to cumulate the individual sales revenues using different periods, and using characteristics such as Products and Customers.
Example of the difference between non-cumulative and cumulative key figures:
Sales volume (cumulative value):
Sales volume 01.20 + sales volume 01.21 + sales volume 01.23 gives the total sales volume for these three days.
Warehouse stock (non-cumulative key figure):
Stock 01.20 + stock 01.21 + stock 01.23 does not give the total stock for these three days.
Technically, non-cumulatives are stored using a marker for the current time (current non-cumulative) and the storage of non-cumulative changes, or inflows and outflows. The current, valid end non-cumulative (to 12.31.9999) is stored in the marker. You can determine the current non-cumulative or the non-cumulative at a particular point in time. To do this, you use the current, end non-cumulative and the non-cumulative changes and/or the inflows and outflows.
Queries for the current non-cumulative can be answered very quickly, since the current non-cumulative is created as a directly accessible value. There is only one marker for each combination of characteristic values. This is always updated when the non-cumulative InfoProvider (InfoProvider that contains the non-cumulative key figures) is compressed. So that access to queries is as quick as possible, compress the non-cumulative InfoProviders regularly. (See Compressing InfoCubes) to keep the marker as current as possible.
In month 03 for example, the marker is read with three non-cumulative changes for a query. In month 04, the marker is updated so that the current marker has to be read with only one non-cumulative change for a query in month five. If the marker were not updated, it would still have four non-cumulative changes to read.
Data Transfer or Storage, and Aggregation for Non-Cumulative Key Figures
To optimize the data transport and data retention for non-cumulative key figures in the BW system, non-cumulative key figures are treated differently from cumulative values in both technical data transfer and storage:
Non-cumulative key figures are mapped using one key figure for non-cumulative changes or two key figures for inflows and outflows.
More information: Non-Cumulative Key Figures.
A non-cumulative InfoCube has to contain a time-reference characteristic, meaning that there must be a time-reference characteristic for exception aggregation of the non-cumulative key figure.
More information: Time Reference Characteristics.
A non-cumulative key figure always has a time-related exception aggregation.
More information: Aggregational Behavior of Non-Cumulative Key Figures.
In certain cases it might be necessary to define the validity of a non-cumulative.
More information: Validity Area.
Non-cumulatives are transferred in an initialization run and in the change runs that follow (initialization can also be omitted here).
More information: Transferring Non-Cumulative Data into BW.
In query definition and navigation, cumulative and non-cumulative key figures are handled in exactly the same way. Cumulative and non-cumulative key figures can be analyzed at the same time in a query.
Special handling of non-cumulative key figures is only possible in InfoCubes and DataStore objects (advanced).
If you are using a SAP HANA database, note that non-cumulative processing with reference points is only supported in certain releases. You can check the release using method IS_REFP_UPDATE_SUPPORTED in class CL_RSDRC_NCUM_PROCESSOR. If your SAP HANA database supports non-cumulative processing with reference points, all new InfoCubes with non-cumulative key figures will be processed in this way. If you have InfoCubes that were created with an earlier SAP HANA release, you can convert them using report RSDD_SWITCH_NCUM_HANDLING.