[
https://issues.apache.org/jira/browse/SPARK-61838?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Hyukjin Kwon resolved SPARK-61838.
----------------------------------
Fix Version/s: connect-rust-4.4.0
Resolution: Fixed
Issue resolved by pull request 709
[https://github.com/apache/spark-connect-rust/pull/709]
> Reduce peak memory while collecting rows
> ----------------------------------------
>
> Key: SPARK-61838
> URL: https://issues.apache.org/jira/browse/SPARK-61838
> Project: Spark
> Issue Type: Bug
> Components: Connect
> Affects Versions: connect-rust-4.4.0
> Reporter: Lavan Vivekanandasarma
> Assignee: Lavan Vivekanandasarma
> Priority: Major
> Labels: pull-request-available
> Fix For: connect-rust-4.4.0
>
>
> The Rust-backed Python client retains the entire native result vector while
> constructing Python Rows. Large string and binary payloads therefore remain
> alive in both native and Python representations during collection, causing
> avoidable peak memory use.
> The proposed change consumes native rows by value and releases each native
> payload after its Python Row is constructed. It applies to collect,
> first/head, take and tail, preserving returned values, Row behavior and
> public method signatures. Iterator and configuration-read behavior remain
> unchanged.
--
This message was sent by Atlassian Jira
(v10.3.26#10030026)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]