Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!ames!haven!vrdxhq!daitc!daitc.daitc.mil
From: jkrueger@daitc.daitc.mil (Jonathan Krueger)
Newsgroups: comp.databases
Subject: Re: How do you...
Message-ID: <472@daitc.daitc.mil>
Date: 2 Apr 89 22:51:12 GMT
References: <1707@anasaz.UUCP> <466@daitc.daitc.mil> <514@cullsj.UUCP>
Sender: jkrueger@daitc.daitc.mil
Reply-To: jkrueger@daitc.daitc.mil (Jonathan Krueger)
Organization: Defense Applied Information Technology Center
Lines: 18
In-reply-to: gupta@cullsj.UUCP (Yogesh Gupta)

John Moore wrote:
>> >A further restriction is that it must be in alphabetical order.

I proved that I didn't read his article carefully:
>> But you said you didn't care if they
>> were sorted, you just wanted no extra cost for it.

Yogesh Gupta spotted my mistake.  Thanks to Yogesh, apologies to John.

As noted by others, your query optimizer should decide to perform the
select first, the sort on the selected rows second.  If it's really
good it might skip the sort if the storage structure maintains an
order.  Failing that, a single pass to verify that it's sorted should
cost little.  Failing even that, you can skip the explicit "order by"
requirement and depend on side effects of the storage structure.

-- Jon
--