Re: initial value on create table/index statement
From: John Hurley <johnbhurley_at_sbcglobal.net>
Date: Mon, 29 Jun 2009 12:24:27 -0700 (PDT)
Message-ID: <3f357ff5-d20e-4f83-a32d-9daa679c98cf_at_a36g2000yqc.googlegroups.com>
On Jun 29, 3:14 pm, John Hurley <johnbhur..._at_sbcglobal.net> wrote:
> On Jun 26, 5:46 pm, DEV <sdp..._at_gmail.com> wrote:
>
> snip
>
>
>
> > > > Is there a way to change the value of "inital" in "create table" or
> > > > "create index" during imp? Pre-creating table is not an option here.
> > > > I am taking a export dump with rows=n from production and importing it
> > > > into dev/QA and since the production tables/indexes have larger values
> > > > for "inital", it takes a lot of storage. Is there a way to create
> > > > tables/indexes with smaller inital values during import?
> > > > Thanks,
> > > > Dev
>
> > > What oracle versions are you going to/from in the export/import
> > > process?
>
> > > How are the relevant tablespaces defined in the to/from process?
>
> > Both are on same version (10.2.0.3) and defined the same way.
> > EXTENT MANAGEMENT LOCAL AUTOALLOCATE
> > BLOCKSIZE 8K
> > SEGMENT SPACE MANAGEMENT AUTO
>
> You might consider uniform extent sizing in the target tablespace
> (s) ... been a while now since I played with import in that
> situation. Any reason you are staying away from impdp?
>
> Another possible approach and much more flexible is to use
> dbms_metadata to extract out schema based ddl ... I do a bunch of
> stuff that way. Then you can get individual files created per table
> if you wish and push as much as you want into some kind of offline
> editing/fixing process.
Date: Mon, 29 Jun 2009 12:24:27 -0700 (PDT)
Message-ID: <3f357ff5-d20e-4f83-a32d-9daa679c98cf_at_a36g2000yqc.googlegroups.com>
On Jun 29, 3:14 pm, John Hurley <johnbhur..._at_sbcglobal.net> wrote:
> On Jun 26, 5:46 pm, DEV <sdp..._at_gmail.com> wrote:
>
> snip
>
>
>
> > > > Is there a way to change the value of "inital" in "create table" or
> > > > "create index" during imp? Pre-creating table is not an option here.
> > > > I am taking a export dump with rows=n from production and importing it
> > > > into dev/QA and since the production tables/indexes have larger values
> > > > for "inital", it takes a lot of storage. Is there a way to create
> > > > tables/indexes with smaller inital values during import?
> > > > Thanks,
> > > > Dev
>
> > > What oracle versions are you going to/from in the export/import
> > > process?
>
> > > How are the relevant tablespaces defined in the to/from process?
>
> > Both are on same version (10.2.0.3) and defined the same way.
> > EXTENT MANAGEMENT LOCAL AUTOALLOCATE
> > BLOCKSIZE 8K
> > SEGMENT SPACE MANAGEMENT AUTO
>
> You might consider uniform extent sizing in the target tablespace
> (s) ... been a while now since I played with import in that
> situation. Any reason you are staying away from impdp?
>
> Another possible approach and much more flexible is to use
> dbms_metadata to extract out schema based ddl ... I do a bunch of
> stuff that way. Then you can get individual files created per table
> if you wish and push as much as you want into some kind of offline
> editing/fixing process.
Sorry guess I should have added that using some plsql and utl_file along with dbms_metadata provide you with tools for slicing and dicing schema based definitions any way you please. Received on Mon Jun 29 2009 - 14:24:27 CDT