You are viewing a plain text version of this content. The canonical link for it is here.
Posted to issues@spark.apache.org by "Cheng Lian (JIRA)" <ji...@apache.org> on 2014/05/16 13:23:36 UTC
[jira] [Commented] (SPARK-1368) HiveTableScan is slow
[ https://issues.apache.org/jira/browse/SPARK-1368?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=13998619#comment-13998619 ]
Cheng Lian commented on SPARK-1368:
-----------------------------------
Corresponding PR: https://github.com/apache/spark/pull/758
> HiveTableScan is slow
> ---------------------
>
> Key: SPARK-1368
> URL: https://issues.apache.org/jira/browse/SPARK-1368
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Reporter: Michael Armbrust
> Fix For: 1.1.0
>
>
> The major issues here are the use of functional programming (.map, .foreach) and the creation of a new Row object for each output tuple. We should switch to while loops in the critical path and a single MutableRow per partition.
--
This message was sent by Atlassian JIRA
(v6.2#6252)