Re: Fastest way to create new table based on a view?

On Wed, 31 Jan 1996, Greg Dora wrote:

> So,
> What does UNRECOVERABLE do? Keep rollback segments from
> being written?
>
> gdora_at_pts.mot.com
>

well...rb is not written to anyways in a "create table as select ..." since its a DDL and hence cannot be rolled back. however, using unrecoverable will also skip redo generation. the option is also available on indexes. there are recovery consequences you should be aware of...check the manaul for details

jc

> In article <4ep892$2sj_at_hpcc48.corp.hp.com> you wrote:
> : Clive Bostock (clive_at_shimmer.demon.co.uk) wrote:
> : : In article <1771AD275S86.PGSCHMID_at_bcsc02.gov.bc.ca>,
> : : PGSCHMID_at_bcsc02.gov.bc.ca writes
> : : >I'd like to create a new table based on a view which contains a join
> : : >of several tables. What is the fastest way to perform this without
> : : >writing the view rows to an OS file and using SQL*Loader to load the
> : : >new table?
> : : >
> : : >So far the SQL*Net COPY seems to be fastest. The new table
> : : >will contain approx 500,000 rows. We are running a read-only Data
> : : >Warehouse environment so I don't care if the new table is recoverable.
> : : >
> : : >Thanks.
> : : >Petr Schmidt
> : : >pgschmid_at_bcsc02.gov.bc.ca
 

> : : Have you tried:
 

> : : CREATE TABLE New_Tab
> : : AS SELECT * FROM View_Name;
 

> : : --
> : : Clive Bostock
> :
> : if you are using 7.2+, you should look into using the UNRECOVERABLE option
> : along with the create table statement.
> :
> : jc
>
  Received on Thu Feb 01 1996 - 00:00:00 CET