Tom Lane wrote:
> "Florian G. Pflug" <fgp@phlo.org> writes:
>> I just stumbled over the following behaviour, introduced with 8.3,
>> and wondered if this is by design or an oversight.
>
> No, this was in 8.2.
Ah, sorry - I'm porting an app from 8.1 straight to 8.3, and blindly
assumes that i'd have worked with 8.2...
>> If you define a domain over some existing type, constrain it to
>> non-null values, and use that domain as a field type in a table
>> definition, it seems to be impossible to declare pl/pgsql variables
>> of that table's row type. The problem seems to be that upon
>> declaration, the row variable is filled with nulls - but since the
>> domain is marked not-null, that immediatly triggers an exception.
>
> What else would you expect it to do? AFAICS any other behavior would
> be contrary to spec.
It's the inconsistency between row types (where the not-null contraint
in the table definition *doesn't* prevent a declaration like "myvar
mytable" in pl/pgsql), and domains (where the not-null constraint *does*
prevent such a declaration) that bugs me.
Plus, the fact that we don't support "default" specifications in
pl/pgsql for row types turns this inconvenience into a major PITA,
forcing you to use "record" when you know that correct type perfectly
well...
Is there some difficulty in implementing row-type defaults, or is it
just that nobody cared enough about them to do the work?
regards, Florian Pflug
