[
https://issues.apache.org/jira/browse/HIVE-14404?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=15407341#comment-15407341
]
Marta Kuczora commented on HIVE-14404:
--------------------------------------
The Super CSV library is used to generate the output if the outputformat is
dsv. Unfortunately Super CSV doesn’t support multiple characters as delimiter.
They have an open issue for this:
https://github.com/super-csv/super-csv/issues/10
I can think of two possible ways to solve this issue.
1. Implement the change in Super CSV and get a new version of it. However I
don’t know anything about the lifecycle of Super CSV releases, so getting a new
version might take some time.
2. Implement a new OutputFormat in Hive for the dsv output format which
supports the multi-character delimiters. This could use the same logic as
currently if the delimiter is only one character, so only the multiple
character case should be implemented.
Which one would be the preferred way to go?
> Allow delimiterfordsv to use multiple-character delimiters
> ----------------------------------------------------------
>
> Key: HIVE-14404
> URL: https://issues.apache.org/jira/browse/HIVE-14404
> Project: Hive
> Issue Type: Improvement
> Reporter: Stephen Measmer
> Assignee: Marta Kuczora
>
> HIVE-5871 allows for reading multiple character delimiters. Would like the
> ability to use outputformat=dsv and define multiple character delimiters.
> Today delimiterfordsv only uses on character even if multiple are passes.
> For example:
> when I use:
> beeline>!set outputformat dsv
> beeline>!set delimiterfordsv "^-^"
> I get:
> 111201081253106275^31-Oct-2011
> 00:00:00^Text^201605232823^2016051968232151^201605232823_2016051968232151_00000_0_1
>
> Would like it to be:
> 111201081253106275^-^31-Oct-2011
> 00:00:00^-^Text^-^201605232823^-^2016051968232151^-^201605232823_2016051968232151_00000_0_1
>
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
