Date: Mon, 10 Apr 2006 11:44:55 -0700
Reply-To: jfh@stanfordalumni.org
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: Jack Hamilton <jfh@STANFORDALUMNI.ORG>
Subject: Re: Efficiency of Proc sort with out =
In-Reply-To: <BAY101-F37718D3D427D9E86215C89DECC0@phx.gbl>
Content-Type: text/plain; charset=iso-8859-1
Do you remember where you saw that?
Considering some of the code I've seen produced by SAS Institute, I
wouldn't consider them the final experts on efficient SAS code. As you
said earlier, you'd have to test it to be sure.
--- toby dunn <tobydunn@hotmail.com> wrote:
> Jack ,
>
> The reason I say that using out= would be more efficient is that I
> was
> reading a SAS thingy on thier web site and using teh out= was one of
> the
> tips they said would help reduce the I/O.
>
>
>
> Toby Dunn
>
>
>
>
>
> From: Jack Hamilton <jfh@alumni.stanford.org>
> Reply-To: jfh@alumni.stanford.org
> To: toby dunn <tobydunn@HOTMAIL.COM>, SAS-L@LISTSERV.UGA.EDU
> Subject: Re: [SAS-L] Efficiency of Proc sort with out =
> Date: Mon, 10 Apr 2006 10:21:38 -0700 (PDT)
>
> I'd be surprised if it makes much difference. If PROC SORT works
> like
> other procs, it creates a temporary output data set and then renames
> it
> to the final name after it's sure everything worked.
>
> Deleting the old data set probably takes some resources, so creating
> a
> new data set with OUT= might be slightly faster, but I would expect
> that overwriting an existing data set would take about the same
> amount
> of time regardless of its name.
>
> That said, PROC SORT is one of the most heavily researched and
> optimized parts of the SAS system, and it's hard to predict what
> might
> trigger an optimization that results in a noticeable change in run
> time.
>
>
> --- toby dunn <tobydunn@HOTMAIL.COM> wrote:
>
> > Sarah ,
> >
> > Efficient can mean many things to many people. It is highly
> likely
> > that you
> > are both right. To do the arguement justice you will first have
> to
> > define
> > what you mean bby efficiency and then the rest is an academic
> problem
> > of
> > running a sort with a large enough data set to accurately test the
> > two
> > methods. Now having said that I would say that the use of an out=
> > will be
> > more efficient if you are looking at speed. And If one is looking
> at
> > work or
> > permenant disk space and not counting scratch files I would say
> the
> > sort
> > that overwrites.
> >
> >
> >
> > Toby Dunn
> >
> >
> >
> >
> >
> > From: Sarah Gollin <SarahGollin@HBOSPLC.COM>
> > Reply-To: Sarah Gollin <SarahGollin@HBOSPLC.COM>
> > To: SAS-L@LISTSERV.UGA.EDU
> > Subject: Efficiency of Proc sort with out =
> > Date: Mon, 10 Apr 2006 11:41:50 -0400
> >
> > Hi,
> >
> > Just trying to settle an argument. My colleague believes that it
> is
> > more
> > efficient to use an out = statement on a proc sort as opposed to
> > overwriting the original dataset. I have googled this but have
> been
> > unable
> > to come up with a definite answer. Can you help?
> >
> > Thanks,
> >
> > Sarah
> >
>
>
> ---
> Jack Hamilton
> Sacramento, California
>
>
>
>
| 
