To: "Jim C. Nasby" <jnasby@pervasive.com>
cc: Dan Gorman <dgorman@hi5.com>, mischa@ca.sophos.com,
	pgsql-performance@postgresql.org
Subject: Re: Selects query stats? 
In-reply-to: <20060523184100.GW64371@pervasive.com> 
References: <9C4E40CF-6C06-4297-AF32-A7EF2A3F3C0F@hi5.com>
	<4473512E.7020805@ca.sophos.com>
	<82252882-2A15-4823-A517-7AB3161889D1@hi5.com>
	<4473553C.1010804@ca.sophos.com>
	<127D68F6-7620-4A62-8D1A-9B475E7571CD@hi5.com>
	<20060523184100.GW64371@pervasive.com>
Comments: In-reply-to "Jim C. Nasby" <jnasby@pervasive.com>
	message dated "Tue, 23 May 2006 13:41:00 -0500"
Date: Tue, 23 May 2006 15:13:13 -0400
Message-ID: <17713.1148411593@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

"Jim C. Nasby" <jnasby@pervasive.com> writes:
> On Tue, May 23, 2006 at 11:33:12AM -0700, Dan Gorman wrote:
>> In any other DB (oracle, mysql) I know how many queries (selects) per  
>> second the database is executing. How do I get this
>> number out of postgres?

> You can't. You also can't know how many DML statements were executed
> (though you can see how many tuples were inserted/updated/deleted), or
> how many transactions have occured (well, you can hack the last one, but
> it's a bit of a mess).

Hack?  We do count commits and rollbacks (see pg_stat_database); doesn't
seem that hacky to me.

Counting individual statements would add overhead (which the OP already
declared unacceptable) and there are some definitional issues too, like
whether to count statements executed within functions.

			regards, tom lane