> > The project web site has a lot of info, but here is a quick
> > example of what it does:
> > thebook=# show statistics * from db history;
> > db | minutes | Questions | Connections | Com_delete
> > ---------+----------------+-----------+-------------+------------
> > thebook | 10/26/09 09:45 | 1 | 0 | 0
> > thebook | 10/26/09 09:44 | 8 | 1 | 6
> > thebook | 10/26/09 09:43 | 0 | 1 | 0
> We already have various statistics views. They don't cover exactly what
> you are doing here (aggregate by user and host), but it would be simpler
> to extend and augment them instead of introducing a completely new
> > We certainly welcome your feedback and thoughts on this.
> > There is still some hard work to be done. We have an INFORMATION
> > SCHEMA implementation for MySQL 5.x, but not yet for PostgreSQL.
> The information schema is defined in the SQL standard. Unless this
> stuff one day appears there, the information schema is the wrong place
> to look.
> The patch itself appears to be licensed under the GPL, which means we
> can't even look at it.
I had taken a look at existing statistics reporting (at
least in the 7.4.x tree) and didn't see a good analog to what this
provides. Part of the key point of this is getting some standard
You may not have had much time to review material at the
project site, http://www.ExtSQL.com/ -- but we would like to see
this as part of the SQL standard. I represent our Company on the
US ANSI (DM32) committee with responsibility for SQL.
We had proposed this as a comment on an earlier version of
the standard. From discussion with some senior committee members,
it would appear the best fit for this type of capability might
be in the INFORMATION SCHEMA part of the standard. We also had
developed the "SHOW STATISTICS" syntax for DB version that don't
support INFORMATION SCHEMA.
I had reviewed statistics reporting in DB2, Oracle, and
SQL server. It was VERY different between implementations, much
seemed geared toward query optimization (which makes sense) and
tends to be implementation specific.
As you can see, the goal here is reporting on activity
at the SQL level.
Regarding the GPL limitation. That will probably be
removed on code we release to PostgreSQL -- hadn't really thought
about the fact it would be a show stopper. Thanks for bringing
John Murtari Software Workshop Inc.
jmurtari(at)thebook(dot)com "software that fits!" (TM)
(315) 944-0999 (x-211) http://www.SoftwareWorkshop.com/
In response to
pgsql-hackers by date
|Next:||From: Josh Berkus||Date: 2009-10-31 17:37:24|
|Subject: Re: \d+ for long view definitions?|
|Previous:||From: Josh Berkus||Date: 2009-10-31 17:35:22|
|Subject: Re: Patch set under development to add usage reporting.|