Selecting the Components of the Analysis Process

On the Overview tab page, you can make settings for the data source, data analysis and data target.

1. The data source can be an InfoProvider, a database table or a SAP HANA analysis process.
   Data source SAP HANA Analysis Process allows you to execute a series of analysis processes in sequence and only to execute the whole result.
2. The data analysis can be one of the following:
   • Function or Script: A number of selected PAL functions from SAP HANA are available to you here. You can also add your own PAL functions.
     Note To use the PAL functions, you have to install SAP HANA SPS06 with the Application Function Library (AFL). The Application Function Library (AFL) contains the PAL functions. For more information, see the SAP HANA Installation Guide
     Note For your own L functions, you have to implement the following interfaces:

     • IF_RSDHAM_OPER design time specific
     • IF_RSDHAR_OPER runtime specific
     • IF_RSDHAR_OPER_LFUNC L-Language specfic
     • IF_RSDHAF_PARAM_VAL UI specific

     For more information, see the interface documentation in the BW system.

     The following PAL functions are supported:

     • ABC Analysis
     • Apriori
     • Double Exponential Smoothing
     • K-means
     • Outliers
     • Single Exponential Smoothing
     • Triple Exponential Smoothing
     • Weighted Score Table
     You can find the documentation for the PAL functions in the SAP HANA Predictive Analysis Library (PAL).
   • Procedure:
     • You can select a procedure that you have implemented in the SAP HANA Modeler as a content procedure. These are located in a particular schema and are assigned to a package. To refine the search, you can specify the schema or the package where the procedure is located. For a content procedure, a repository procedure is always generated. The analysis process access these repository procedures.

       The advantage of implementing a procedure yourself is that you can reuse it later. You always need to take care of the transport yourself.

     • You can generate a suggestion implementation that you have to adjust to your requirements. In the proposal implementation, the fields are projected, but not modified.

       It is generated in accordance with the assignment information that you define on the source and target. A name is generated according to the following pattern: 0BW:DAP:<Name of Analysis Process>~PN. When the analysis process is activated, the procedure in the repository is generated as a repository procedure in the schema defined for this. This does not have a package assignment and can be found via the schema. The advantage of a generated procedure is that it is transported with the SAP HANA analysis process.
     • You can select an ABAP managed database procedure that you have created in ABAP in Eclipse. The ABAP runtime environment creates a corresponding database procedure on the SAP HANA database. You can then use this database procedure in the SAP HANA analysis process. For more information about ABAP Managed Database Procedures, see the ABAP key word documentation at . http://help.sap.com/abapdocu_740/en/index.htm?file=abenamdp.htm

       You should normally use formulas, especially for calculations by field. The system supports this better. If you use procedures, you need to implement a lot of things yourself that are actually already in the system, currency translation for example. You can create formulas directly on the field, in the details for the data source or data target.

       You should use procedures if you need complete algorithms that should not be calculated by field and for cases that no formulas exist for.

     • You can generate an ABAP Managed Database Procedure and then adapt it to suit your particular requirements. Generated procedures are useful if you do not want to perform calculations by field, and for complex algorithms.

       If you generate an ABAP Managed Database Procedure, the class is generated first of all. This comprised the prefix HAP_ and a generated ID. If the analysis process is changed, the procedure is generated again. You can only insert your own coding in the body. This is not overwritten if the procedure is regenerated. This allows you to enhance the existing method without creating one of your own. If you edit the SAP HANA analysis process from the ABAP Development Tools, you can call the editor directly by double-clicking on the generated procedure. If you edit the SAP HANA analysis process from the BW system, double-clicking on the generated procedure only takes you to the Class Builder. You can display the generated coding here, but cannot edit it.

3. The data target can be one of the following:
   • Analytic Index: You can either have the system generate an analytic index (default setting) or choose an existing one. If you select a generated analytic index as the data target, you need the authorization for creating, changing, loading and deleting data on an analytic index. These authorizations are assigned using authorization object S_RS_AI.

     The result can be saved or, if using the analysis process as the data source for a further analysis process, not saved. If a query is executed on an analytic index, the analysis is executed at runtime if the data is not saved. The control parameters can be entered when executing the query.

   • DataStore Object (classic) for Direct Update
   • Database table
   • Embedded in Data Transfer Process: If you choose this option, you can create a transformation for your SAP HANA analysis process, and thus use this analysis process as a source for other ones, by choosing  Environment  Transformation .
4. Save and activate your SAP HANA analysis process.