From: | Lee Hachadoorian <lee(dot)hachadoorian(at)gmail(dot)com> |
---|---|
To: | Simon Riggs <simon(at)2ndquadrant(dot)com> |
Cc: | pgsql-general <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: Large Rows |
Date: | 2011-10-26 15:11:43 |
Message-ID: | CANnCtnKAc6bkUzibsxd4CxH9iQgSGyMsX7DCHrGJfJyjqMj35w@mail.gmail.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
On Wed, Oct 26, 2011 at 10:15 AM, Simon Riggs <simon(at)2ndquadrant(dot)com> wrote:
> On Wed, Oct 26, 2011 at 2:57 PM, Lee Hachadoorian
> <lee(dot)hachadoorian(at)gmail(dot)com> wrote:
>
> > Interesting. Although your example of one, 100000-dimension array works,
> > five hundred 2-dimension arrays does not work. I can do the SELECT, but
> the
> > CREATE TABLE fails:
> >
> > ERROR: row is too big: size 9024, maximum size 8160
> > SQL state: 54000
> >
> > David has already hit the nail on the head in terms of this being a
> > "political" problem rather than a technology problem. I'm open to ideas,
> but
> > I realize there might be no other answer than "No one in their right mind
> > should do this."
>
> No, this is a technology problem.
>
> Toast pointers are 20 bytes per column, so with 500 columns that is
> 10000 bytes - which will not fit in one block.
>
> If you wish to fit this in then you should use a 2 dimensional array,
> which will then be just 1 column and your data will fit.
>
> --
> Simon Riggs http://www.2ndQuadrant.com/
> PostgreSQL Development, 24x7 Support, Training & Services
>
Very useful to know. Thank you.
--Lee
--
Lee Hachadoorian
PhD, Earth & Environmental Sciences (Geography)
Research Associate, CUNY Center for Urban Research
http://freecity.commons.gc.cuny.edu/
From | Date | Subject | |
---|---|---|---|
Next Message | Mike Blackwell | 2011-10-26 17:26:20 | Re: All and ANY |
Previous Message | Tom Lane | 2011-10-26 14:42:19 | Re: pglesslog for Postgres 9.1.1 |