Re: SQL/MED - core functionality

On Tue, Dec 14, 2010 at 10:48 AM, Itagaki Takahiro
<itagaki(dot)takahiro(at)gmail(dot)com> wrote:
> On Tue, Dec 14, 2010 at 23:45, Robert Haas <robertmhaas(at)gmail(dot)com> wrote:
>>> We need RULEs or INSTEAD OF TRIGGERs to support updatable foreign tables.
>>
>> We do?  Why can't the support for updating foreign tables be built-in
>> rather than trigger-based?
>
> Do we have any concrete idea for the built-in update feature?
> There are no definitions in the SQL standard about interface for updates.
>
> So, I think RULE and TRIGGER are the best solution for now.
> In addition, even if we support some kinds of built-in update feature,
> I still think RULE and TRIGGER are useful, for example, logging purpose.

I think triggers are useful. I see no reason to support rules. If
the first version of our SQL/MED functionality is read-only, that's
fine. But triggers are slow, clumsy, and expose implementation
details to users, so those should be something that we provide as a
way of making the database extensible, not something we use to build
core functionality.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company