Home | About | Sematext search-lucene.com search-hadoop.com
 Search Hadoop and all its subprojects:

Switch to Threaded View
Sqoop, mail # user - Sqoop Halts - Exporting from HDFS to Oracle (Date/Timestamp issue?)


Copy link to this message
-
Re: Sqoop Halts - Exporting from HDFS to Oracle (Date/Timestamp issue?)
Jarek Jarcec Cecho 2013-11-02, 01:25
Hi Tanzir,
would you mind sharing with us the following information?

* Sqoop command that you are using
* Entire generated log when executed with --verbose parameter

Jarcec

On Wed, Oct 30, 2013 at 10:28:41PM +0600, Tanzir Musabbir wrote:
> I'm encountering an issue when exporting data from HDFS to Oracle. I'm not sure whether this is an existing issue or not. When I carefully checked the log, I saw one of the column has "2/13/2013 4:35:50" wheres the corresponding column type in Oracle is varchar2.
>
> Can't we export date or time just as string? When my HDFS has that kind of data, Sqoop halts and at the end its(task) killed by mapreduce task's timeout.
> Based on this: http://qnalist.com/questions/31561/sqoop-modifies-the-date-format-in-the-exported-data, I'm getting impression that Oracle driver converts it to timestamp and tries to insert it and when it doesn't see timestamp as column type it halts. Please correct me if I'm wrong.
> Is there any way not to do that kind of conversion? and just export them as string?
>
>
> Thanks in advance.
>
> Sincerely,Tanzir