To: "Tony Wasson" <ajwasson@gmail.com>
cc: "Jim C. Nasby" <jnasby@pervasive.com>, "Jorge Godoy" <jgodoy@gmail.com>,
	pgsql-general@postgresql.org
Subject: Re: [SQL] (Ab)Using schemas and inheritance 
In-reply-to: <6d8daee30605231754s58364cd1o79a6520f203a1734@mail.gmail.com> 
References: <200605231716.58784.jgodoy@gmail.com>
	<20060523232609.GP64371@pervasive.com>
	<6d8daee30605231754s58364cd1o79a6520f203a1734@mail.gmail.com>
Comments: In-reply-to "Tony Wasson" <ajwasson@gmail.com>
	message dated "Tue, 23 May 2006 17:54:09 -0700"
Date: Tue, 23 May 2006 23:08:58 -0400
Message-ID: <25040.1148440138@sss.pgh.pa.us>
From: Tom Lane <tgl@sss.pgh.pa.us>

"Tony Wasson" <ajwasson@gmail.com> writes:
> $.02 about lots of schemas.

> I worked with an application that had 500 schemas and that worked very
> well. However, as the number of schemas exceeeded 8000 the query speed
> started to degrade. Running \d with a single schema in your search
> path took a few seconds with that many schemas. Queries that were
> running in 100ms were now taking about 600ms.

Note that this probably says more about \d than about anything else.
I can believe that having a large number of schemas listed in your
search path would suck, but there's not a good reason for lots of
schemas unrelated to your session to cause you any great pain.
(\d does some searching that might be impacted by lots of schemas,
but that doesn't say ordinary queries would be.)

If you've got a counterexample please file a bug with details ...

			regards, tom lane