7.3 dataflow modeller: What do I need to know?
The dataflow modeller is a brilliant new function available in the new release of SAP Business Warehouse. It really helps with making your BW system easier to manage, develop and work in.
Specifically, the functions it provides are…
- Top-down modelling of new dataflows
- Organization for existing dataflows
- Creation of template dataflows using Best Practice
- SAP delivered predefined dataflows
That sounds easy! What are you going to tell me?
Yeah, the above sounds straightforward. And using the dataflow modeller is generally easy enough to use.
But due to the complexity of the topic in hand, and the flexibility of the system, there are quite a few things that can catch you out. I struggled to get going with the tool because I was missing a few critical pieces of information and therefore understanding.
Here are some important things you need to know and understand. If you remember these, then you will be able to pick up the dataflow modeller with ease.
1) Nodes within a dataflow are either “persistent” or “non-persistent”
A “persistent node” references an object within the database, such as an InfoCube or DSO. A non-persistent node is one that only exists within the dataflow and does not reference anything within the database. It is, in effect, an empty placeholder for an object, which will be created at some point in the future.
2) There are two types of dataflow: template and standard
Template: A template dataflow only contains a structure and therefore is only permitted to contain non-persistent nodes. This is really a dataflow that is created for use as a guide to create other dataflows. You would use it as a pattern when creating a standard dataflow.
Standard: A standard dataflow contains nodes that reference objects existing in the database. Both non-persistent and persistent nodes may be contained within this type of dataflow.
You can create a standard dataflow, either from scratch, or use a template dataflow as a starting point.
3) Non-persistent nodes can be made persistent within a dataflow
If you create a dataflow with non-persistent nodes, it will look nice but won’t actually reference anything within the database – not very useful when building a data warehouse.
Within a dataflow modeller, you can change a non-persistent node into a persistent node. By using the right-click contextual menu, you can create a new object that the node will reference. Or, as an alternative, use the node’s contextual menu to configure the node so that it references an object that already exists.
4) Persistent Nodes can be dragged from the Admin Workbench (RSA1) into the dataflow modeller
The dataflow modeller makes extensive use of drag-and-drop. You can include existing objects, such as InfoCubes or DSOs, within a dataflow by dragging them from an AWB panel on the left of your screen and placing them into your datamodel. You may need to use the “show navigator panel” button, above the main window, to get the AWB to appear.
5) A dataflow can also be navigated via a hierarchical navigator
SAP has provided two ways to navigate a single dataflow: the flowchart, or the hierarchical layout. Both these views contain the same nodes and provide alternative views of the same data.
Within the AWB you have a new tab, sensibly titled “Dataflows”. Find your dataflow, expand the hierarchy and you will see all the persistent nodes organized by type. Note that only non-template dataflows can be navigated in this way. This is because template dataflows can contain only non-persistent nodes, which are not shown within the hierarchical view.
Figure 1: Hierarchical view of a dataflow
Figure 2: Dataflow modeller view of a dataflow