On Mon, Sep 29, 2003 at 12:14:17PM -0400, Hardy Merrill wrote:
> Tim Bunce [EMAIL PROTECTED] wrote:
> > On Mon, Sep 29, 2003 at 09:32:10AM -0400, Hardy Merrill wrote:
> > > Thomas A. Lowery [EMAIL PROTECTED] wrote:
> > > > # Does the error message contains the phrase: duplicate key
> > > > if ($dbh->errstr =~ m/duplicate\skey/i) {
> > > ^^^^^^^^^^^^^^
> > > Just to clarify, as Thomas said above, each database produces
> > > different error messages to indicate that you are trying
> > > to insert a duplicate key - so what you search for here will
> > > vary depending on which database you are using.
> >
> > Of course, in an ideal world, DBD's would all set $DBI::state
> > ($h->state) to the appropriate value: "23000"
> >
> > Maybe someone could define a list of the errors that most frequently
> > need to be explicitly checked for. The DBI spec could then list
> > those along with the standard SQL_STATE error codes and suggest that
> > drivers support at least that basic set of codes.
>
> I think this is an excellent idea!
Does that mean you're volunteering to coordinate making the list?
That is, soliciting input from dbi-users and dbi-announce and produce
a list of the corresponding SQL_STATE values (based on ODBC 3.5).
I hope so :)
Tim.
> The only situation I can think of that I've had to code
> around where different databases produce different results
> is this one - this "trying to insert a duplicate record"
> error. There must be others ;-)
>
> --
> Hardy Merrill
> Red Hat, Inc.
