Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!watmath!clyde!rutgers!mit-eddie!genrad!decvax!ucbvax!UORDBV.BITNET!JCOV
From: JCOV@UORDBV.BITNET.UUCP
Newsgroups: mod.computers.vax
Subject: Britton Lee problem
Message-ID: <8702092256.AA24492@ucbvax.Berkeley.EDU>
Date: Mon, 9-Feb-87 07:17:00 EST
Article-I.D.: ucbvax.8702092256.AA24492
Posted: Mon Feb  9 07:17:00 1987
Date-Received: Wed, 11-Feb-87 05:11:27 EST
Sender: wold@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 40
Approved: info-vax@sri-kl.arpa


 
I have a question about VMS used with the Briton Lee
data base machine.
 
 
   We have omnibase v5.0 and right now the copyin command is out
to lunch so I wrote a program to do the equivalent thing
i.e. read a file and do insert sql statements to get some
data accross.
 
(1) I find I get about o I/O's per second (11-780).
Its not usiing up much VAX cpu time, but it takes
forever to get the job done.  I wonder why insert is so slow
compared to copyin? If there is some legitimate reason
for this, can I do the same thing that copyin does -- even
if it is some programming, if it will speed it up by
a factor of 50 or so which seems to be the case, then its worth it.
I am using PL/I and RSQL interfact -- should I be using
precompiler and a supported  language? if so,
can I do whatever the precompiler does and fake it out?
 
   The I/O's I am talking about are direct I/O's which seems
to mean that it takes about 8 I/O's for each (small) record to go
accross the parallell interface.  I have seen programs
such as the one to install the software on the Britton Lee
do things much faster than that.
 
   After I was all finished, I tried to create a unique
index and got a duplicate key message.  Is there
anyway at that point to tell what the key is or should
I have created the index before inserting all those
records and had the program print the key?
 
   Any assistance would be appreciated.
 
 
   Thanks in advance.
 
         John Covici