[
https://issues.apache.org/jira/browse/SPARK-4042?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14179939#comment-14179939
]
Yin Huai commented on SPARK-4042:
---------------------------------
Can you also add some test results? Like the amount of data read from HDFS when
we read all columns and only a single column.
> append columns ids and names before broadcast
> ---------------------------------------------
>
> Key: SPARK-4042
> URL: https://issues.apache.org/jira/browse/SPARK-4042
> Project: Spark
> Issue Type: Bug
> Components: SQL
> Affects Versions: 1.1.0
> Reporter: wangfei
>
> appended columns ids and names will not broadcast because we append them
> after creating table reader. This leads to the config broadcasted to executor
> side dose not contain the configs of appended columns and names.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
