Re: Inserting 300.000.000 datasets ..

From: Hans-Jürgen Schönig <a9601497_at_unet.univie.ac.at>
Date: 2000/07/15
Message-ID: <dBXb5.41758$7D2.739504@news.chello.at>#1/1

Im Artikel <396f048e.2526972_at_news-server> schrieb nsouto_at_nsw.bigpond.net.au.nospam (Nuno Souto):
> On Fri, 14 Jul 2000 08:55:47 GMT, "Hans-Jürgen Schönig"
> <a9601497_at_unet.univie.ac.at> wrote:
>

>>I have a process that generates about 300.000.000 datasets which I need
>>to import into an Oracle 8.1.6 on RedHat Linux. Since I don't have the
>>space to store and then insert the data I'd like to do this via a pipe
>>(or named pipe). Is there any possibility to use SQL-Loader? I tried
>>named pipes but  all my attempts failed.
>>

>
> Named pipes should work. What you need is a program that reads the
> datasets and pumps them out the pipe WITHOUT closing the output file
> unit at the end of each dataset. You'll prolly have to write a C
> program to do that, I don't think any of the UNIX standard filters will
> do it. As soon as a EOF makes it thrpugh to the named pipe, SQL*Loader
> will get it on the other side and promptly terminate.
>
> Cheers Nuno Souto nsouto_at_nsw.bigpond.net.au.nospam
> http://www.users.bigpond.net.au/the_Den/index.html

I thought that named pipes should work and I have already some experience with named pipes but I had absolutely no chance - i'd simpl didn't work. Have you ever tried named pipes with SQL-Loader? Is there any possibility to use ordinary pipes?

        Hans Received on Sat Jul 15 2000 - 00:00:00 CDT