Steve Atkins <steve(at)blighty(dot)com> writes:
> IF TG_OP='INSERT' OR NEW.foo != OLD.foo THEN
> -- do stuff
> Is there anything that guarantees that the NEW.foo != OLD.foo won't
> be evaluated (and hence fail, as there is no OLD) on an insert?
The only way is to split the expression into two IF commands. The real
problem is not with the evaluation per se, but with plpgsql having to
pass down all the variable values used in the expression before
evaluation can start.
regards, tom lane
In response to
pgsql-general by date
|Next:||From: Rosny||Date: 2004-12-20 20:24:17|
|Subject: Installing postgres 8 using Remote Desktop Connection|
|Previous:||From: Tom Lane||Date: 2004-12-20 19:53:40|
|Subject: Re: Diskspace |