Creating Process Chains

Now, you have created models for product, customer, and sales data, then you have filled tables and used the data in higher-level views and reports. But, loading data manually for many models would be too much work. Therefore, you want to automatize this process. SAP Datasphere SAP BW bridge offers an object type that specifies a sequence of run-time steps such as loading or deleting data: the process chain. In contrast to a simple task list, a process chain can define which tasks run in parallel.

In this lesson, let’s delve into creating, managing, and scheduling process chains and the related terminology:

• Process type: Each process step belongs to exactly one process type and each type belongs to a process category, based on its function. For example, a process chain can contain two different processes of the process type Data transfer process from category "Loading process and postprocessing". It defines the tasks that the process will perform and the required parameters you have to provide during maintenance.
• Process variant: A process variant is a stored set of parameters for a specific process type, for example, it specifies a specific DTP. A variant is useful because it allows you to reuse the specific configuration of a process when defining a new one.
• Process instance: This is the actual occurrence of the process when the chain is running.

To understand the relationship between these terms, imagine that you are cooking a meal. The process types are similar to the dishes, such as pumpkin soup, lasagna, or ice cream. To make it easier to find what you need, they are grouped into categories such as appetizer, side dish, main dish, and dessert. Typically, you want to make sure that the appetizer is served first, then the main course with the side dishes in parallel, and dessert after everything else is eaten. The process variant is like the version of the recipe that you want to follow, for example, it determines which ice cream flavor you use and how many scoops you serve. The instance is like a cooking event - each time you use a different pot, and the duration is slightly different.

The process type and variant allow for customization and flexibility in your data management tasks. They are the building blocks of a process chain. The instance is the key for traceability and accountability, ensuring you can accurately track and log each step within your chain for auditing and troubleshooting purposes.

Let’s have a look at a few important process types:

Start Process: The start process defines the start conditions of a process chain. Every chain needs one start process at its beginning, like turning on the electricity is necessary before you start cooking. The same start process cannot be used in other process chains.

You must define a start process with one of the following conditions:

• Immediate start upon scheduling the chain
• Start at a specified time, similar to setting an alarm clock.
• Event: Start after a particular event.
• It can be raised using an API based on an event from a source.
• It can be triggered by another process chain, known as a meta chain.

Collector Process: If parallel paths are defined in a process chain, a collector process ties them together and defines in which case a subsequent process is started. The following collector processes exist:

• AND (last): All predecessor tasks must be completed and the subsequent job starts when the last predecessor is finished.
• OR (each): Each completed predecessor task triggers an instance of the subsequent process.
• XOR (first): Any predecessor process must be completed and only the first finished predecessor job triggers the subsequent process.

Launch the following video to learn how you create a process chain.

To sum up, here are the steps:

1. Select 'Create' to begin a new process chain. You are greeted with a canvas in the graphical editor, representing the space where you’ll map out your chain.

2. Insert Processes: Use buttons like Insert Process, Insert Collector, and Insert Decision to add different types of nodes to your canvas.

3. Connect Nodes: Draw connections between nodes to establish the flow of processes. For most of the nodes, you can specify if the successor should run:
   • only in case of error
   • only in case of success
   • or in all cases

4. Define Details of the Processes: For each node, specify the process type and parameters. One parameter is typically the name of the object and other parameters are performance settings or limits.

5. Use the Properties button to define properties for the process chain, for example the job priority.
6. Save, activate, and schedule the process chain then start the monitor.