Re: [HACKERS] [PATCH] Lockable views

On Fri, 27 Oct 2017 07:11:14 +0200
Robert Haas <robertmhaas(at)gmail(dot)com> wrote:

> On Wed, Oct 11, 2017 at 11:36 AM, Yugo Nagata <nagata(at)sraoss(dot)co(dot)jp> wrote:
> > In the attached patch, only automatically-updatable views that do not have
> > INSTEAD OF rules or INSTEAD OF triggers are lockable. It is assumed that
> > those views definition have only one base-relation. When an auto-updatable
> > view is locked, its base relation is also locked. If the base relation is a
> > view again, base relations are processed recursively. For locking a view,
> > the view owner have to have he priviledge to lock the base relation.
>
> Why is this the right behavior?
>
> I would have expected LOCK TABLE v to lock the view and nothing else.
>
> See http://postgr.es/m/AANLkTi=KupesJHRdEvGfbT30aU_iYRO6zwK+fwwY_sGd@mail.gmail.com
> for previous discussion of this topic.

This discussion is one about 7 years ago when automatically-updatable views
are not supported. Since 9.3, simple views can be updated as well as tables,
so now I think it is reasonable that LOCK TABLE for views locks their base
tables.

If we want to lock only the view, it seems to me that LOCK VIEW syntax is good.
However, to realize this, changing the syntax to avoid a shift/reduce
conflict will be needed as disucussed in the "LOCK for non-tables" thread.

>
> --
> Robert Haas
> EnterpriseDB: http://www.enterprisedb.com
> The Enterprise PostgreSQL Company
>
>
> --
> Sent via pgsql-hackers mailing list (pgsql-hackers(at)postgresql(dot)org)
> To make changes to your subscription:
> http://www.postgresql.org/mailpref/pgsql-hackers

--
Yugo Nagata <nagata(at)sraoss(dot)co(dot)jp>