Re: loading tools (followup)

Have you tried writing this in Pro*C or PRO*Cobol and validating all the data and doing array inserts/updates? How much data is coming in each night? How bad is the performance, and is all the sql optimized?

"Egon" <egonvdb_at_xs4all.nl> wrote in message news:8v2qbq$b3v$1_at_news1.xs4all.nl...
> The REPLACE option does a truncate of the table first, this is not an
 option
> for me, I an using Oracle 7.3.2.2.0 on SunOS 5.6. I have tried to make
 some
> sql
> and pl/sql scripts and it is possible to get what I want but the time it
> takes
> to make the updates and insert is a major problem. Right now I am dealing
> with
> 1 database and about 30 tables and the number of records in them wil grow
 on
> a
> daily bases
>
>
>
> Eric Givler <egivler_at_flash.net> schreef in berichtnieuws
> Ac%Q5.3208$6W1.217924_at_news.flash.net...
> > With that in mind, I know there is a REPLACE option with SQL*Loader, but
 I
> > think that does a TRUNCATE on the table first prior to inserting the
 rows.
> > That would not be a good thing. I think the other option is the APPEND
 to
> > add stuff on the end. I haven't used it in quite a while, so I haven't
 read
> > any documentation on it lately.
> >
> > Can you tell me which version you are using and on what platform? I can
> > look it up tomorrow and see if we might be missing something...
> >
> > If it does NOT have the option we are looking for, why not just load it
 into
> > ANOTHER table (a temporary table), then write a pl/sql routine that
 loads
 it
> > into the real table and does the inserts or updates depending on the
 keys.
> > You could add MORE validation to that and produce a report with either
> > DBMS_OUTPUT OR UTL_FILE. For things like this, I usually write
 PRO*cobol
> > load routines, but I'm from a VAX/VMS background...
> >
> > "Egon" <egonvdb_at_xs4all.nl> wrote in message
> > news:8v0l9h$g74$1_at_news1.xs4all.nl...
> > > Is there a tool available to load ASCII data into tables and if a
> > > record(key) exists
> > > update that record.
> > >
> > >
> >
> >
>
>
Received on Sat Nov 18 2000 - 04:51:34 CET