On Thu, Dec 07, 2006 at 02:52:11PM -0800, Alan Hodgson wrote:
> On Thursday 07 December 2006 08:38, "Angva" <angvaw@gmail.com> wrote:
> > three commands. For instance I have a hunch that creating the indexes
> > first (as I do now) could slow down the clustering - perhaps the row
> > locations in the indexes all have to be updated as the cluster command
> > shifts their locations? And perhaps vacuuming should be done before
> > clustering so that dead tuples aren't "in the way"?
>
> clustering also removes the dead tuples.
>
> I would just:
>
> - create one index, the one to be clustered
> - cluster the table
> - create the remaining indexes
And then run ANALYSE. No need to vacuum because the cluster did that
already.
Have a nice day,
--
Martijn van Oosterhout <kleptog@svana.org> http://svana.org/kleptog/
> From each according to his ability. To each according to his ability to litigate.