Why not use import ... BUT
1. drop all indices (to be rebuild later - you would (IMO) want to do this
regardless of imp/SQLLOADER)
2. drop all key constraints (again enabling after import)
3. disable logging on the table while it's being loaded via imp therefore
generating no redo work
4. import the table (indexes=n constraints = n); I totally agree that
SQLLOADER is faster than imp, but exp/imp is a helluva lot easier to execute
5. re-enable logging on the table and other things like indices and
constraints
Just a thought
Jim Stivers
"Marcotullio, Angelo" wrote:
> Orlando wrote: "There's a way to bypass rollback- and index creation during
> the import, which speeds things up a lot. I believe this is the same
> technique used by the SQL*Loader "direct path" option."
>
> No, this is not correct. SQL*Loader "direct path" writes database blocks
> directly. It does not do inserts. Any form of import will do inserts.
>
> -----Original Message-----
> From: Orlando Andico [mailto:[EMAIL PROTECTED]]
> Sent: Friday, June 08, 2001 11:50 AM
> To: Marcotullio, Angelo
> Cc: [EMAIL PROTECTED]
> Subject: RE: Data Extract
>
> On Fri, 8 Jun 2001, Marcotullio, Angelo wrote:
> ..
> > Oracle export (exp) creates a binary file that can only be loaded using
> the
> > import (imp) program. The problem with import is that it uses insert
> > statements to load the data.
>
> There's a way to bypass rollback- and index creation during the import,
> which speeds things up a lot. I believe this is the same technique used by
> the SQL*Loader "direct path" option.
>
> --
> Orlando Andico <[EMAIL PROTECTED]>
> Mosaic Communications, Inc.
>
> -----BEGIN GEEK CODE BLOCK-----
> Version: 3.1
> GE d(-) s: a-25 C++++ UBLSI++++$ P+++ L+++>++++ E- W++ N(+)
> o K? w O-- M- !V PS(++) PE- Y PGP-- t(+)@ 5(+) X++@ R(+) tv@
> b++ DI++ G e++@ h--(*) r% y+
> ------END GEEK CODE BLOCK------
