Date:         Wed, 8 May 2002 09:29:31 -0600
Reply-To:     JGerstle@SW.UA.EDU
Sender:       "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From:         John Gerstle <JGerstle@SW.UA.EDU>
Subject:      Revised stupid question
In-Reply-To:  <5.1.0.14.2.20020508150039.03524ec0@pop3.powernet.co.uk>
Content-type: text/plain; charset=US-ASCII

Actually, I might be wrong with the list of PROCs I mentioned.  It's
still early and my coffee hasn't woken me up completely yet.  I
know that there are some PROCs, maybe MEANS, TTEST, plus
some others.  It's almost second nature for me to add PROC
SORT before any analysis PROC just to prevent that error
message from coming up, regardless if that PROC needs to be
sorted or not.
John

> At 08:54 08/05/02 -0600, John Gerstle wrote (in part):
>
> >As we know, there are some analysis PROCs that will give you error
> >messages that the data was not sorted on the CLASS variable (Ie GLM,
> >ANOVA, TTEST, etc).
>
> John, at risk of making a fool of myself (yet again :-), I have to say
> that this is a new one on me.  Indeed, I've always considered one of
> the benefits of the CLASS statement (particularly when BY processing
> is an alternative) is that is does not require the CLASS variable(s)
> to be sorted.  Can you give, or point to, example(s) of what you
> describe?
>
> Kind Regards,
>
>
> John
>
> ----------------------------------------------------------------
> Dr John Whittington,       Voice:    +44 (0) 1296 730225
> Mediscience Services       Fax:      +44 (0) 1296 738893
> Twyford Manor, Twyford,    E-mail:   John.W@mediscience.co.uk
> Buckingham  MK18 4EL, UK             mediscience@compuserve.com
> ----------------------------------------------------------------