Introduction

Ruletypes and Rulegroups not only define operational related definitions but also value based definitions of a transformation.

The operational related definition of the rulegroups in the transformation focuses on how the data is processed and how the datarecords go from their source structure into the target structure. Value based definitions of the transformation then actually determine the fieldvalue in the target structure.

Value based and operational related definitions

Every datarecord coming from the source structure runs through the start- and endrountine of a transformation. (Not necessarily only once!) Each of this process of the data is defined as a rulegroup.

The standard rulegroup is existing as default and it can’t be deleted but additional ones can be created. For each rulegroup there is an additional datarecord in the target structure.

The result of all rulegroups is merged in the target structure of the transformation and then posted to the target objects. (InfoCube etc.)

The number of records not necessarily multiplies by the number of rulegroups defined as the records get filtered by the identical keys. Key are created on the criteria of the  target structure/object by the system.

The key fields can be seen in the transformation with a key symbol

The graphical display also shows based on which input parameters the target fieldvalue is derived from. For example if a summation is done from two keyfigures of the source to one keyfigure on the target. Two arrows from the source fields merge into the target field. Fieldvalues can be derived through:

• Direct Assignment
• Formula
• No tranformation
• Routine
• Routine with Unit

additionally character infoobjects can:

• Read from MasterData
• Read From DataStore
• Initial

All ruletypes also handle conversion exits if defined. 

Dataflow example with two rulegroups

The example shows a split of amount between two sales groups with the help of a second rulegroup as described above.

This kind of data manipulations should only happen in the entry or datamart layer and it’s not recommended in the datawarehouse or transformationlayer as the records should be passed along without data manipulations. (Aside from deletion in start or endroutine).

The result of all rulegroups is merged in the target structure of the transformation and then posted to the target objects. (InfoCube etc.)