You are viewing a plain text version of this content. The canonical link for it is here.

Posted to issues@spark.apache.org by "Matei Zaharia (JIRA)" <ji...@apache.org> on 2014/10/20 22:49:34 UTC

[jira] [Updated] (SPARK-3467) Python BatchedSerializer should dynamically lower batch size for large objects

     [ https://issues.apache.org/jira/browse/SPARK-3467?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel ]

Matei Zaharia updated SPARK-3467:
---------------------------------
    Assignee: Davies Liu

> Python BatchedSerializer should dynamically lower batch size for large objects
> ------------------------------------------------------------------------------
>
>                 Key: SPARK-3467
>                 URL: https://issues.apache.org/jira/browse/SPARK-3467
>             Project: Spark
>          Issue Type: Improvement
>          Components: PySpark
>            Reporter: Matei Zaharia
>            Assignee: Davies Liu
>             Fix For: 1.2.0
>
>
> If you try caching largish objects in Python, you will get a crash sooner than you would in Scala / Java because Python automatically batches them. I believe the default batch size is 10, though it may be 1024. But maybe we can start by pickling the first object and using a smaller batch size if it is large.



--
This message was sent by Atlassian JIRA
(v6.3.4#6332)

---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscribe@spark.apache.org
For additional commands, e-mail: issues-help@spark.apache.org