| From: | Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
|---|---|
| To: | Gregory Stark <stark(at)enterprisedb(dot)com> |
| Cc: | Heikki Linnakangas <heikki(dot)linnakangas(at)enterprisedb(dot)com>, Simon Riggs <simon(at)2ndQuadrant(dot)com>, Postgres <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Our CLUSTER implementation is pessimal |
| Date: | 2008-09-09 12:22:14 |
| Message-ID: | 22854.1220962934@sss.pgh.pa.us |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Gregory Stark <stark(at)enterprisedb(dot)com> writes:
> In particular I'm thinking of people clustering on a covering index (which
> isn't as uncommon as it sounds, if you have a covering index you probably do
> want to cluster it -- consider many-to-many join tables). We should be able to
> do an index-only scan which might be even faster than sorting.
[ scratches head... ] You need *all* the data from the heap. Or by
"covering index" do you mean an index that contains the entire table
contents? Doesn't really sound like a case we need to focus on; or
at least this version of clustering isn't what it needs, it wants an
implementation where the table and the index are the same thing.
regards, tom lane
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Tom Lane | 2008-09-09 12:24:04 | Re: Synchronous Log Shipping Replication |
| Previous Message | Simon Riggs | 2008-09-09 12:21:14 | Re: Synchronous Log Shipping Replication |