Return-Path: mao
Received: by postgres.Berkeley.EDU (5.61/1.29)
	id AA12523; Tue, 23 Mar 93 09:59:53 -0800
Message-Id: <9303231759.AA12523@postgres.Berkeley.EDU>
From: Mike Olson <mao@postgres.Berkeley.EDU>
Subject: Re: Opposite of 'addattr' command ?
To: postgres@postgres.berkeley.edu
Sender: pg_adm@postgres.berkeley.edu
In-Reply-To: Your message of Mon, 22 Mar 1993 20:23:19 PST.
             <9303230423.AA09593@postgres.Berkeley.EDU> 
Date: Tue, 23 Mar 1993 10:00:29 -0800
From: Mike Olson <mao@postgres.Berkeley.EDU>

In message <9303230423.AA09593@postgres.Berkeley.EDU>you write:

> Do I miss something or is there no opposite of the
> 'addattr' command ? I would like to be able to remove
> attributes from a class. Is there a way to do this ?
> (except copying the whole thing without the attribute marked
> for deletion, destroying the class, and renaming the copy) ?

attribute deletion isn't supported.  you can do a 'retrieve into' to
create a new table without the attributes you don't want.

> The magical 8K limit: Is it valid for a single attribute, the
> whole class, or both ? Can the limit be raised by changing the
> include/port/?/machine.h definition of BLCKSZ, or is this more
> complicated ?

redefining BLCKSZ almost certainly won't do the trick.  the limit is
for a single tuple; if you have an attribute that's very large, you
should consider using large objects to store it.  see section 7 of
the manual.
					mike
