[
https://issues.apache.org/jira/browse/SPARK-7323?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Sean Owen updated SPARK-7323:
-----------------------------
Description: Currently we invoke combiners.insert() for each tuple in
Aggregator - which results in creation of an Iterator per tuple, and iterating
over this iterator : instead, we can directly call insertAll to avoid the
object creation, method invocation and iteration overhead - for each tuple when
combiners are used. (was:
Currently we invoke combiners.insert() for each tuple in Aggregator - which
results in creation of an Iterator per tuple, and iterating over this iterator
: instead, we can directly call insertAll to avoid the object creation, method
invocation and iteration overhead - for each tuple when combiners are used.)
Priority: Minor (was: Major)
> Use insertAll instead of individual insert while merging combiners
> ------------------------------------------------------------------
>
> Key: SPARK-7323
> URL: https://issues.apache.org/jira/browse/SPARK-7323
> Project: Spark
> Issue Type: Improvement
> Components: Spark Core
> Reporter: Mridul Muralidharan
> Assignee: Mridul Muralidharan
> Priority: Minor
> Fix For: 1.3.2, 1.4.0
>
>
> Currently we invoke combiners.insert() for each tuple in Aggregator - which
> results in creation of an Iterator per tuple, and iterating over this
> iterator : instead, we can directly call insertAll to avoid the object
> creation, method invocation and iteration overhead - for each tuple when
> combiners are used.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
