Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: INSERT into HUGE table

Re: INSERT into HUGE table

From: Roger Palmen <r.palmen.ani_at_alewijnse.nl>
Date: 2000/08/11
Message-ID: <3993BC56.8D0E943F@alewijnse.nl>#1/1

Well, what would be the most efficient way, besides using SQLLoader? This job has to run on user command, and I hate having things in a system that are OS and / or Installation dependent, which happens when you step outside the database.
Isn't there any way of doing an replace in an insert? Maybe using an INSTEAD-OF INSERT trigger?

Thanks for your help.

hermannkall_at_my-deja.com wrote:

> Hi Roger,
>
> use SQL*Loader in conventional mode and all records having
> duplicate keys will be rejected.
> Make sure that BINDSIZE and ROWS is set properly to avoid the default
> commit after every 64 records.
>
> cheers
> Hermann
>
> In article <39927C92.9F93ADE1_at_alewijnse.nl>,
> Roger Palmen <r.palmen.ani_at_alewijnse.nl> wrote:
> > Hi all,
> >
> > I'm trying to find the most efficient way to insert large amounts of
> > records (lets say 100K a day) into a single huge table (100M records).
> > However, the problem is that the set of records to insert contains
> > records already in the target table. These duplicate records can
 either
> > be overwritten or neglected.
> >
> > A simple but effective INSERT statement fails if only one record
 doesn't
> > pass the PK-constraint. What is the most efficient way to do this?
> >
> > - Export / SQL-loader?
> > - Delete rows from source?
> > - Use PL/SQL and cursor to insert one record at a time?
> > - Outer join with target?
> > - etc?
> >
> > Using Oracle 8.1.6 on NT.
> >
> > Thanks for advice.
> > Please reply to newsgroup, or pe1pvy_at_amsat.org NOT the sender.
> >
> >
>
> Sent via Deja.com http://www.deja.com/
> Before you buy.
Received on Fri Aug 11 2000 - 00:00:00 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US