spark-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Nicholas Hwang (JIRA)" <j...@apache.org>
Subject [jira] [Updated] (SPARK-9021) Change RDD.aggregate() to do reduce(mapPartitions()) instead of mapPartitions.fold()
Date Sun, 19 Jul 2015 13:43:04 GMT

     [ https://issues.apache.org/jira/browse/SPARK-9021?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Nicholas Hwang updated SPARK-9021:
----------------------------------
    Summary:  Change RDD.aggregate() to do reduce(mapPartitions()) instead of mapPartitions.fold()
 (was: Have pyspark's RDD.aggregate() make a deepcopy of zeroValue for each partition)

>  Change RDD.aggregate() to do reduce(mapPartitions()) instead of mapPartitions.fold()
> -------------------------------------------------------------------------------------
>
>                 Key: SPARK-9021
>                 URL: https://issues.apache.org/jira/browse/SPARK-9021
>             Project: Spark
>          Issue Type: Bug
>          Components: PySpark
>    Affects Versions: 1.4.0
>         Environment: Ubuntu 14.04 LTS
>            Reporter: Nicholas Hwang
>
> Please see pull request for more information.
> I initially "patched" this arguably unexpected behavior by serializing zeroValue, but
ended up mimicking the deepcopy approach used by other RDD methods. I also contemplated having
fold/aggregate accept "zero value generator functions" instead of an actual object, but that
obviously changes the API.
> Looking forward to hearing back and/or being educated on how I'm inappropriately using
this functionality (relatively new to Spark and functional programming). Thanks!



--
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


Mime
View raw message