Re: Trigger Question with Commit

From: hpuxrac <johnbhurley_at_sbcglobal.net>
Date: Sat, 9 Aug 2008 14:45:36 -0700 (PDT)
Message-ID: <d966940b-24d1-4870-9606-76c05d0ca73d@k7g2000hsd.googlegroups.com>


On Aug 9, 6:37 am, Tim X <t..._at_nospam.dev.null> wrote: snip

> > Triggers are complicated and have always had the possibility of
> > introducing complications and side effects.  Many people in oracle
> > including Tom Kyte think that so many people cannot code and test them
> > accurately that they should be eliminated.
>
> > Why post questions in an internet news group if you aren't going to
> > include all the code and just describe your impression of what is
> > going on?
>
> > Much recommended ... build your own test cases.  Check it out
> > yourself.  Read Tom Kyte's books!

> Good advice. Also, the OPs conclusion that its an implicit commit sounds
> vary dubious. As far as I know, DDL is the only place you will get an
> implicit commit (apart from some clients, which 'implicitly' commit,
> such as Perl's DBI. However, its been a while since I did any Oracle
> development (8i) and I'm only now getting back into things, so I could
> be mistaken.

Words like "implicit commit" make me nervous. In general most of the time I find that people that combine words like that just don't understand the oracle transaction model, or plsql procedures, etc. Received on Sat Aug 09 2008 - 16:45:36 CDT

Original text of this message