[
https://issues.apache.org/jira/browse/SQOOP-1522?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14224946#comment-14224946
]
Jarek Jarcec Cecho commented on SQOOP-1522:
-------------------------------------------
I believe that it's reasonable to expect that our users will want to define the
resulting structure of imported files (what separators, quote characters should
be used, ...). That is pretty much in line what we have in Sqoop 1 with
parameters such as {{--separated-by}}, {{--enclose-by}} and others already. So
I'm +1 on providing this functionality to end user.
I would prefer to keep this in the HDFS connector though as this is specific to
output (input) formatting. I would prefer to keep the CSV IDF as strict as
possible as that is not a place to do transformations, it's a layer for
exchanging data between two connectors. The less variability we will allow
there the less issue we will see when one connector will do something slightly
different than the other connector expects.
> Sqoop2: User defined separators and quote character
> ---------------------------------------------------
>
> Key: SQOOP-1522
> URL: https://issues.apache.org/jira/browse/SQOOP-1522
> Project: Sqoop
> Issue Type: Sub-task
> Components: sqoop2-hdfs-connector
> Affects Versions: 1.99.3
> Reporter: Abraham Elmahrek
>
> IDF currently uses CSV with ',' field separator, '\n' row separator, and ' as
> a quote character. It would be great to have the HDFS connector output be
> user defined.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
