unclear docs on https://googlecloudplatform.github.io/google-cloud-python/latest/pubsub/#publishing

[ensonic]

The docs have this example:

publisher = pubsub.PublisherClient()
topic = 'projects/{project_id}/topics/{topic}'.format(
 project_id=os.getenv('GOOGLE_CLOUD_PROJECT'),
 topic='MY_TOPIC_NAME', # Set this to something appropriate.
)
publisher.create_topic()
publisher.publish(topic, b'My first message!', spam='eggs')

should it be:

publisher = pubsub.PublisherClient()
topic_path = 'projects/{project_id}/topics/{topic}'.format(
 project_id=os.getenv('GOOGLE_CLOUD_PROJECT'),
 topic='MY_TOPIC_NAME', # Set this to something appropriate.
)
topic = publisher.create_topic(topic_path)
publisher.publish(topic, b'My first message!', spam='eggs')

main questions:

• how does publisher.create_topic() know what topic if we don't tell it
• is there still a a topic object as there was before 0.28.0?
• can we just create a topic regardless whether it is already there? The old code (pre 0.28.0) was doing:

client = pubsub.Client()
topic = client.topic(topic_name)
if not topic.exists():
 topic.create()

[tseaver]

How does publisher.create_topic() know what topic if we don't tell it?

Thanks for pointing this out: Client.create_topic definitely requires the name argument.

Is there still a a topic object as there was before 0.28.0?

No, the redesigned API dropped the Topic, Subscription, and Snapshot classes.

Can we just create a topic regardless whether it is already there?

No: that would raise a Conflict exception.

[ensonic]

Thanks for the quick update. So how do I now check if a topic already exists? Create it and handle the exception?