You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@beam.apache.org by "Udi Meiri (JIRA)" <ji...@apache.org> on 2018/08/13 19:54:00 UTC
[jira] [Commented] (BEAM-4275) Pubsub: add DirectRunner support for
id_label and timestamp_attribute in Python SDK
[ https://issues.apache.org/jira/browse/BEAM-4275?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=16578838#comment-16578838 ]
Udi Meiri commented on BEAM-4275:
---------------------------------
DirectRunner doesn't support reading or writing label_ids, nor writing timestamp attributes.
> Pubsub: add DirectRunner support for id_label and timestamp_attribute in Python SDK
> -----------------------------------------------------------------------------------
>
> Key: BEAM-4275
> URL: https://issues.apache.org/jira/browse/BEAM-4275
> Project: Beam
> Issue Type: Bug
> Components: runner-direct, sdk-py-core
> Reporter: Udi Meiri
> Assignee: Udi Meiri
> Priority: Major
>
> At least for publishing (and maybe pulling) messages, non-Dataflow-based sources and sinks for Pub/Sub use the [public API|https://cloud.google.com/pubsub/docs/publisher] for Pub/Sub, which doesn't support id_label and timestamp_attribute settings.
> Publishing:
> id_label - add an attribute to each message with a unique value
> timestamp_attribute - add an attribute to each message with the publishing time as its value
> Pulling:
> id_label - use the value of this message attribute to deduplicate messages
> timestamp_attribute - use the value of this message attribute as the element's timestamp
>
> Implementation details: could probably create a pubsubio.py module, for reuse with other runners (i.e. implement Pub/Sub IO as PTransforms and not NativeSinks and Sources).
--
This message was sent by Atlassian JIRA
(v7.6.3#76005)