Very Computer

. .dra.com!feed1.news.erols.com!cpk-news-hub1.bbnplanet.com!news.bbnplanet.com?
. !dispatch.news.demon.net!demon!netcom.net.uk!not-for-mail

.Newsgroups: comp.databases.informix
Subject: Re: ISQL Begin Work & Commit Work
.Date: Tue, 11 Mar 1997 21:24:49 -0800
.Organization: Iota Data Systems Ltd



.NNTP-Posting-Host: dialup-17-24.netcomuk.co.uk
.Mime-Version: 1.0
.Content-Type: text/plain; charset=us-ascii
.Content-Transfer-Encoding: 7bit
.X-Mailer: Mozilla 3.0Gold (Win16; I)
.Lines: 58
.Xref: uustar.starnet.net comp.databases.informix:43461

Hi, Richard.

Let's assume you're doing a LOAD.

I would execute something like:

BEGIN WORK;
LOCK TABLE this_tab IN EXCLUSIVE MODE; {Optional but I prefer this}
LOAD FROM "this_data.unl" INSERT INTO this_tab;

I would not include a COMMIT or ROLLBACK at this stage.  If the LOAD
runs without crashing and claims to have loaded the number of rows I
expected I would then run COMMIT WORK as a singleton statement or, if
the LOAD failed, run a ROLLBACK WORK.

In practice I might well have split the load file so that actual script
would be:

{COMMIT WORK;}
BEGIN WORK;
LOCK TABLE this_tab IN EXCLUSIVE MODE; {Optional but I prefer this}
LOAD FROM "xaa" INSERT INTO this_tab;

If all went well I'd uncomment the COMMIT and change xaa to xab.

Likewise if I had a long list of operations I'd start with:

BEGIN WORK;
<command for 1st operation>
{
COMMIT WORK;
BEGIN WORK;
<command for 2nd operation>
...
etc
...

Quote:}

and then work my way down commenting out completed operations and
uncommenting the next operation, always starting with the COMMIT for the
previous job.

Regards

Ian