Re: Export Tablespaces

"FlameDance" <FlameDance_at_gmx.de> wrote in message news:bq76af$7e1$01$1_at_news.t-online.com...
> Niall Litchfield wrote:
>
> > The point is that if all
> > datafiles are 2gb and the table is larger than 2gb the create table
> > statement in the import *will* fail.
>
> That'll be my next worry. Regardless of compress=y or compress=n? Or
> does you proposal of compress=n solve the issue?

yes. Compress=y means recalculate the extent sizes so that the entire table fits in one extent. Hence the table will try to allocate more than the size [Quoted] [Quoted] of any single datafile as the initial extent. This will fail. compress=n doesn't change initial and next so you end up with more smaller extents that can span the datafiles.

>
> > Actually if the table is 1.7gb and
> > isn't in a tablespace on its own and isn't the first object imported the
> > import is already likely to fail for lack of 1.7gb contiguous space to
> > create the initial extent.
>
> I can enforce that by importing only that table in a first import run.
> Problem is, on the next import run with all objects I get millions of
> unique key violations on that table, even with ignore=y set.
>
> Can I exclude a certain object from an import run?

[Quoted] from memory....

tables=(table list)

> If yes, how?
> If not, can I suppress the unique key violation messages?
>
> Stephan
>
>
Received on Fri Nov 28 2003 - 13:05:17 CET