Re: SQL load or DBLOAD

Posted in 1998

Art S. Kagel — Wed, 24 Jun 1998 11:41:25 -0400 — source: Informix-list mailing list archive (1991-1998)

Joe Lumbley wrote: > > Carlson@WHSmith (carlson1@bellsouth.net) wrote: > : Informix: 7.23.UC4 > : OS: HPUX 10.01 > > <snip> > > Go with the dbload. orders of magnitude difference I agree and would just add: Use split to break the input into N files and run N copies of dbload -r, one for each file obviously, this scales very well. I've done this with at least 2X NUMCPUVPS processes with incremental throughput gains for even the last process. Art S. Kagel