Re: Impdp - Full Database
Date: Sun, 4 Jul 2010 07:48:12 -0700 (PDT)
Message-ID: <9ca6a9cd-1986-4d7b-b02d-a12d919cb023_at_x27g2000yqb.googlegroups.com>
On Jul 2, 6:43 pm, Bill <billshatne..._at_googlemail.com> wrote:
> Hi there.
>
> I'm using impdp to import a full database.
>
> Everything seemed to go well, except that the sequences didn't seem to
> match the old db.
>
> There are numerous schema owners and lots of dependencies.
>
> I was hoping to bring the database across in one go but I was
> wondering if anyone can recommend the steps for this situation.
>
> My old method with exp/imp was:
>
> 1. full exp.
> 2. imp rows=n and constraints=n (to build structure)
> 3. disable triggers
> 4. imp rows=y
> 5. imp indexes=y constraints=y
>
> It worked but it took ages.
>
> Any suggestions? I have looked at the Oracle 10g utilities
> documentation but couldn't see anything specific to this.
>
> Thanks, Ed.
Actually posting the sequence information for a specific sequence as it appears in the old db and as it appears in the new db might allow someone to provide more specific information. but If the sequences do not exist in hte target db then the import task would create the sequences in the target. The sequence values should more or less 'match' the source. I say more or less cause the values assigned will skip the buffered values in the old database.
On the other hand if the sequences already exist in the target then the import will not replace them and the values will reflect the existing values rather than the values you need for the freshly imported data. This can happen when you import a production extract into test without removing all the target objects, sequences in this case, from test first.
HTH -- Mark D Powell -- Received on Sun Jul 04 2010 - 09:48:12 CDT