[
https://issues.apache.org/jira/browse/SPARK-12997?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15160270#comment-15160270
]
Hyukjin Kwon commented on SPARK-12997:
--------------------------------------
If I got this correctly, I think the issue itself is a little bit questionable
at the end. If we can apply the same rules for casting, then it would be great
to re-use the {{Cast()}} but both CSV and JSON data sources are dealing with
data types differently.
As long as we are not going to change this behaviour for backwards
compatibility, I believe this issue is invalid or this will end up with some
additional logics like I just did in the PR above with some cases to deal with
the differences.
> Use cast expression to perform type cast in csv
> -----------------------------------------------
>
> Key: SPARK-12997
> URL: https://issues.apache.org/jira/browse/SPARK-12997
> Project: Spark
> Issue Type: Sub-task
> Components: SQL
> Reporter: Reynold Xin
>
> CSVTypeCast.castTo should probably be removed, and just replace its usage
> with a projection that uses a sequence of Cast expressions.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
