- Type Parameters:
T - type of object in the input collection
U - type of object in the output collection
- All Implemented Interfaces:
- PipelineConfigurable, Serializable
public class DynamicSparkCompute<T,U>
extends SparkCompute<T,U>
This class is required to make sure that macro substitution occurs each time a pipeline is run instead of just
the first time the pipeline is run. Without this, the SparkCompute plugin gets serialized into the checkpoint and
re-loaded every subsequent time with the properties from the first run.
- See Also:
- Serialized Form