Re: pg_stat_*_columns?

From: Jim Nasby <Jim(dot)Nasby(at)BlueTreble(dot)com>
To: Joel Jacobson <joel(at)trustly(dot)com>, Pg Hackers <pgsql-hackers(at)postgresql(dot)org>
Subject: Re: pg_stat_*_columns?
Date: 2015-06-05 17:30:42
Message-ID: 5571DCC2.2040106@BlueTreble.com
Views: Raw Message | Whole Thread | Download mbox | Resend email
Thread:
Lists: pgsql-hackers

On 6/5/15 6:51 AM, Joel Jacobson wrote:
>
> 1. I think it would be helpful for DBAs to better understand their own
> system.
> Finding unused *tables* is today easy thanks to pg_stat_*_tables, but
> knowing if something is accessing a *column* or not is not easy.
> In my case all our database access is via sprocs, so I can just grep the
> source code for the column name to see if something is using it, but
> most DBAs probably don't have that luxury.

I have wanted this exact thing when considering vertical partitioning.
It's easy to tell from a size standpoint what columns are good
candidates for putting in a 'side table', but it's very hard to know how
often columns are actually used.

BTW, I think the right way to measure this would be how many rows were
returned for queries referencing a column. Simply knowing how many
queries reference a column doesn't tell you much; you want to know how
much column data was actually pulled out.
--
Jim Nasby, Data Architect, Blue Treble Consulting, Austin TX
Data in Trouble? Get it in Treble! http://BlueTreble.com

In response to

Browse pgsql-hackers by date

  From Date Subject
Next Message Peter Geoghegan 2015-06-05 17:39:42 Re: Further issues with jsonb semantics, documentation
Previous Message Andrew Dunstan 2015-06-05 17:23:37 Re: Further issues with jsonb semantics, documentation