On Fri, Oct 1, 2010 at 4:33 AM, Leonardo Francalanci <m_lists(at)yahoo(dot)it> wrote:
> I guess that if the planner makes a wrong choice in this case (that is,
> seq scan + sort instead of index scan) there's no way for "cluster" to
> behave in a different way. If, on the contrary, the "create table..." uses
> the right plan, and cluster doesn't, we have a problem in the patch.
> Am I right?
Good point. I think you're right.
--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise Postgres Company