You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@beam.apache.org by "Ryan Skraba (JIRA)" <ji...@apache.org> on 2017/04/03 15:54:41 UTC

[jira] [Commented] (BEAM-1847) Read.Unbounded can't specify both max records and max duration.

    [ https://issues.apache.org/jira/browse/BEAM-1847?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15953692#comment-15953692 ] 

Ryan Skraba commented on BEAM-1847:
-----------------------------------

My mistake, it's KafkaIO that ignores `withMaxReadTime()` when `withMaxNumRecords()` is set -- `Read.Unbounded` does permits you combine the stopping conditions.

> Read.Unbounded can't specify both max records and max duration.
> ---------------------------------------------------------------
>
>                 Key: BEAM-1847
>                 URL: https://issues.apache.org/jira/browse/BEAM-1847
>             Project: Beam
>          Issue Type: Bug
>          Components: beam-model
>            Reporter: Ryan Skraba
>            Assignee: Ryan Skraba
>
> Some Beam IOs that expose the ability to turn an unbounded source into a bounded source.  
> For example, KafkaIO uses the underlying Read.from() API to specify the {{withMaxNumRecords}} and/or {{withMaxReadTime}}.  If the former is specified, the latter is silently ignored.  
> I would expect that the first stopping condition to be reached (either max records OR max duration) would stop the source.  
> The underlying implementation {{BoundedReadFromUnboundedSource}} has this logic, but it is not supported -in Read.Unbounded- or the Beam IOs that expose this feature.



--
This message was sent by Atlassian JIRA
(v6.3.15#6346)