On Mon, Dec 13, 2010 at 4:26 AM, Hampus Wessman
<hampus.wessman@gmail.com> wrote:
> CLUSTER should be able to do that for you. See
> http://www.postgresql.org/docs/9.0/static/sql-cluster.html.
>
That's the exact opposite of what he wished to accomplish.
The only thing I can suggest is to do random order of updates, perhaps
while also setting the fill factor for the pages to something large so
there's no room to re-use a page, causing more shuffling.
I suspect you'll have to do this for every table you've got, else you
could leak some information about what rows were updated since the
last import if the person has access to the file used to import the
data originally. You'll also have to do this every time you dump the
DB, I'd suspect...
