Re: sqlldr from pipe

From: Thomas Kyte <tkyte_at_us.oracle.com>
Date: Sat, 21 Jul 2001 21:52:17 GMT
Message-ID: <a4lhitg101tdg03utpm4do20knl2vea688_at_4ax.com>


A copy of this was sent to Nick Taylor <ntwysiwyg_at_dataskill.co.uk> (if that email address didn't require changing) On Wed, 13 Jun 2001 17:28:41 GMT, you wrote:

>Hi
>
>We're using Oracle 8.1.6 under SuSE Linux. Is there any way to pipe data
>into the standard input of sqlldr?
>
>I would like to be able to bulk load tables on a regular basis from
>output from a legacy system dump program, but without dumping all data
>into files first.
>
>Additionally this would allow input from gzip files via:
>"zcat file | sqlldr..."
>
>I am unaware of any alternative loading mechanism that is anywhere near
>as fast as sqlldr using direct load. My java version does my job, but
>relatively slowly, I guess because of the logging. This simply then
>harks back to the numerous requests for loading without logging :-(
>
>TIA
>
>Nick

use a named pipe.

mknod input.dat p

zcat file > input.dat &
sqlldr .... data=input.dat

--
Thomas Kyte (tkyte_at_us.oracle.com) Oracle Service Industries
Howtos and such: http://asktom.oracle.com/ http://asktom.oracle.com/~tkyte/
Oracle Magazine: http://www.oracle.com/oramag

Expert one on one Oracle, programming techniques and solutions for Oracle.
http://www.amazon.com/exec/obidos/ASIN/1861004826/  

Opinions are mine and do not necessarily reflect those of Oracle Corp 
Received on Sat Jul 21 2001 - 23:52:17 CEST

Original text of this message