[
https://issues.apache.org/jira/browse/SPARK-15240?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Liang-Chi Hsieh updated SPARK-15240:
------------------------------------
Summary: Use buffer variables to improve buffer
serialization/deserialization in TungstenAggregate (was: Use buffer variables
for update/merge expressions instead duplicate serialization/deserialization)
> Use buffer variables to improve buffer serialization/deserialization in
> TungstenAggregate
> -----------------------------------------------------------------------------------------
>
> Key: SPARK-15240
> URL: https://issues.apache.org/jira/browse/SPARK-15240
> Project: Spark
> Issue Type: Improvement
> Components: SQL
> Reporter: Liang-Chi Hsieh
>
> We do serialization/deserialization on aggregation buffer in
> TungstenAggregate for each aggregation function. It wastes time on duplicate
> serde for the same grouping keys. We can optimize it.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
