Google-cloud-dataflow: Why pipeline run twice with DirectRunner?

Question

Given the data set as below

{\"slot\":\"reward\",\"result\":1,\"rank\":1,\"isLandscape\":false,\"p_type\":\"main\",\"level\":1276,\"type\":\"ba\",\"seqNum\":4254         


        

Answer 1

The with statement for Pipeline runs the pipeline. Specifically:

with beam.Pipeline(...) as p:
  [...code...]

is equivalent to:

p = beam.Pipeline(...)
[...code...]
p.run().wait_until_finish()

See the implementation.