- Each operator must provide schema information, i.e. what kind of data is created
- Schema has information about:
- The processing type (IStreamObject): e.g. tuple, keyvalueobject, xml
- The schema of the type (typically only for tuples)
- The schema of the meta types (SDFMetaSchema): e.g. TimeInterval, Latency, …
- Some contraints on the schema
- A flag indicating, if the stream is potentially out of order
- Use SDFSchemaFactory to create output schema from input schema (and preserve information that is not changed by the operator)
Examples
- ReOrderAO (Schema same as input, but ordered)
- ToTupleAO (Input schema is another type, output is tuple, keep meta schema)