[
https://issues.apache.org/jira/browse/SPARK-49547?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]
Ruifeng Zheng resolved SPARK-49547.
-----------------------------------
Fix Version/s: 4.1.0
Resolution: Fixed
Issue resolved by pull request 52440
[https://github.com/apache/spark/pull/52440]
> Support returning RecordBatches from applyInArrow
> -------------------------------------------------
>
> Key: SPARK-49547
> URL: https://issues.apache.org/jira/browse/SPARK-49547
> Project: Spark
> Issue Type: Sub-task
> Components: PySpark, SQL
> Affects Versions: 4.0.0
> Reporter: Adam Binford
> Assignee: Adam Binford
> Priority: Major
> Labels: pull-request-available
> Fix For: 4.1.0
>
>
> Currently the new `applyInArrow` method in PySpark uses a function that takes
> a `pyarrow.Table` and returns a `pyarrow.Table`. This limits the ability for
> this function to scale, as the entire result set must fit in memory at once
> in a `Table`. However, we have use cases that can result in a large amount of
> data that needs to be returned from the function on certain edge cases. The
> result is immediately turned into a series of batches from the Table, so
> there's no reason to not just allow an iterator of batches to be returned
> instead.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
