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

Switch to Threaded View
Sqoop, mail # user - Importing from OracleDB


Copy link to this message
-
Re: Importing from OracleDB
Jarek Jarcec Cecho 2013-02-01, 18:52
Hi James,
I'm glad to hear that you found reasonable workaround! If you think that supporting synonyms make sense, please do not hesitate and open new JIRA [1] for that.

Jarcec

Links:
1: https://issues.apache.org/jira/browse/SQOOP

On Fri, Feb 01, 2013 at 06:14:21PM +0000, James Hogarth wrote:
> On 1 Feb 2013 17:27, "Jarek Jarcec Cecho" <[EMAIL PROTECTED]> wrote:
> >
> > Hi James,
> > I'm afraid that using synonyms is not supported by Oracle driver. The
> problem is that Oracle driver is issuing following query to get columns for
> imported table:
> >
> >   SELECT COLUMN_NAME FROM ALL_TAB_COLUMNS WHERE OWNER = ? AND TABLE_NAME
> = ? ORDER BY COLUMN_ID
> >
> > This query returns no rows for synonym and thus import fails.
> >
> > As a workaround I would recommend to try out Generic JDBC Connector that
> is not using this query to fetch table metadata. You can force Sqoop to use
> Generic JDBC Connector by passing "--driver" parameter with value
> oracle.jdbc.OracleDriver, e.g.
> >
> >   sqoop import --connect ... --driver oracle.jdbc.OracleDriver
> >
> > Another solution that might possibly help you is to install OraOop
> connector, but I'm not sure whether it supports synonyms or not.
> >
> > Jarcec
> >
>
> Hi Jarcec
>
> Thanks for confirming that.
>
> In the end my workaround with specifying the columns capitalised ended up
> working nicely and isn't so bad seeing as I needed to cast the oracle type
> 'number' to Long/BIGINT on a couple of the columns anyway...
>
> Cheers
>
> James