dplutils.pipeline.stream.StreamingGraphExecutor.task_resolve_output

dplutils.pipeline.stream.StreamingGraphExecutor.task_resolve_output#

abstractmethod StreamingGraphExecutor.task_resolve_output(pending_task: Any) StreamBatch[source]#

Return a StreamBatch from completed task

This function takes the output produced by either task_submit() or split_batch_submit(), and returns a StreamBatch object which tracks the length of returned dataframe(s) and the object which references the underlying DataFrame.

The data member of returned StreamBatch will be either:

  • passed to another call of task_submit() in a list container, or

  • yielded in the execute() call (which yields in the user-called run method). If any handling must be done prior to yield, implementation should do so in overloaded execute().