Re: Should I CLUSTER on PRIMARY KEY

[Robert James <srobertjames@xxxxxxxxx>]

Thanks, Chris.  Is there a way to do this deterministically, or at least programatically? I have code to create the tables and cluster them automatically?

On Sun, Jul 19, 2009 at 8:21 PM, Chris <dmagick@xxxxxxxxx> wrote:

Robert James wrote:

I would like to CLUSTER a table on its PRIMARY KEY.  Now, I haven't explicitly defined and named an index for this table - but the primary key defines one.   How can I tell Postgres to CLUSTER on it?

Get the index name:

\d tablename

Right at the bottom it will have the index names:

Indexes:
   "a_pkey" PRIMARY KEY, btree (a)


then cluster:

# cluster tablename using a_pkey;
CLUSTER

Also: If I define an index on a PK, will Postgres make a second one, or realize its redundnant?

Depends how you define it (I think). What's your create table statement look like?

--
Postgresql & php tutorials
http://www.designmagick.com/