While trying to reproduce the pipeline using the Python SDK I've noticed that the topic format is different for the write transform [1] compared to the read transform [2]. It seems it uses /topics/<project>/<topic> and /projects/<project>/topics/<topic> respectively. This is also documented in the Python SDK documentation [3].
Although note the doc string for PubSubWritePayload says > // Topic format is: /topics/project_id/subscription_name presumably meaning topic_name rather than subscription_name. I'll try using the different format and report back. [1] https://pkg.go.dev/github.com/apache/beam/sdks/v2/go/pkg/beam/model/pipeline_v1#PubSubWritePayload [2] https://pkg.go.dev/github.com/apache/beam/sdks/v2/go/pkg/beam/model/pipeline_v1#PubSubReadPayload [3] https://beam.apache.org/releases/pydoc/2.34.0/apache_beam.io.external.gcp.pubsub.html
