[
https://issues.apache.org/jira/browse/BEAM-671?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=17546161#comment-17546161
]
Kenneth Knowles commented on BEAM-671:
--------------------------------------
This issue has been migrated to https://github.com/apache/beam/issues/18049
> Update Read PTransform to implicitly use maxNumRecords and maxReadTime
> ----------------------------------------------------------------------
>
> Key: BEAM-671
> URL: https://issues.apache.org/jira/browse/BEAM-671
> Project: Beam
> Issue Type: Improvement
> Components: sdk-java-core
> Reporter: Jean-Baptiste Onofré
> Priority: P3
> Labels: Clarified
>
> Today, almost all IOs {{Read}} {{PTransform}}, when using an
> {{UnboundedSource}}, do the following in the {{apply()}}:
> {code:java}
> org.apache.beam.sdk.io.Read.Unbounded<byte[]> unbounded =
> org.apache.beam.sdk.io.Read.from(getSource());
> PTransform<PBegin, PCollection<byte[]>> transform = unbounded;
> if (maxNumRecords() < Long.MAX_VALUE || maxReadTime() != null) {
> transform =
> unbounded.withMaxReadTime(maxReadTime()).withMaxNumRecords(maxNumRecords());
> }
> return input.getPipeline().apply(transform);
> {code}
> To avoid to duplicate this code bunch of time, it would make sense to do it
> by default in {{org.apache.beam.sdk.io.Read}}.
--
This message was sent by Atlassian Jira
(v8.20.7#820007)
