Every data flow begins with a source object. You create source objects from datastores or flat files.
When you drag and drop metadata from a datastore or a file format onto the data flow workspace, the system generates a source object.
Datastore Source Objects
To read data from a table or another object in a datastore, you must first import its metadata. Once imported, the object is available in the object library, and you can drag and drop it into your data flow as a source.
SAP Data Services analyzes the metadata to define the input data structure, known as the schema. This schema provides the necessary information to continue building the data flow. During execution, the system uses the datastore configuration to connect to the source and retrieve the actual data.
File Format Source Objects
The same principle applies to file formats. However, after creating the source object in the data flow, you can modify specific properties, such as the file location or error handling options.
If a file format was initially created using a sample file on a local drive, you must update it to access files from the Job Server. This update is required because the Job Server executes the job and must have access to the data files.
You can also create multiple source objects from the same file format within a single data flow. This allows you to fetch different files that share the same structure and join them, for example.
Summary
In this lesson, you have learned how to add sources to your data flows.
- Source objects are created by dragging metadata from datastores or flat files into the data flow workspace.
- Datastore sources require imported metadata to define the input schema and enable data retrieval during execution.
- File format sources allow property modifications including file location and error handling after creation.
- File locations must be updated from local drives to Job Server-accessible paths for proper job execution.
- Multiple source objects from one file format enable processing different files with identical structures within a data flow.