[
https://issues.apache.org/jira/browse/FLINK-36209?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Arvid Heise resolved FLINK-36209.
---------------------------------
Fix Version/s: kafka-3.3.0
Assignee: xiaochen.zhou
Resolution: Fixed
Merged into main as 52e7e580cd3b453a92a87afff4daf19ec53212dd.
> Remove redundant operations in the initialization of KafkaSourceEnumState
> -------------------------------------------------------------------------
>
> Key: FLINK-36209
> URL: https://issues.apache.org/jira/browse/FLINK-36209
> Project: Flink
> Issue Type: Improvement
> Components: Connectors / Kafka
> Reporter: xiaochen.zhou
> Assignee: xiaochen.zhou
> Priority: Minor
> Labels: pull-request-available
> Fix For: kafka-3.3.0
>
>
> In certain methods, such as the
> {{DynamicKafkaSourceEnumerator#onHandleSubscribedStreamsFetch()}} method,
> {{partitions}} are divided into {{assignedPartitions}} and
> {{unassignedInitialPartitions}} before being passed as parameters to the
> {{KafkaSourceEnumState}} constructor. However, within the constructor, these
> {{assignedPartitions}} and {{unassignedInitialPartitions}} are recombined
> into {{{}partitions{}}}, leading to unnecessary operations and reduced
> performance. By optimizing the code to pass {{partitions}} directly as a
> parameter when initializing {{{}KafkaSourceEnumState{}}}, we can eliminate
> redundant operations and enhance performance.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
