Date: Wed, 8 Dec 1999 11:01:51 -0800
Reply-To: David Cassell <cassell@MERCURY.COR.EPA.GOV>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: David Cassell <cassell@MERCURY.COR.EPA.GOV>
Organization: OAO Corp.
Subject: Re: Free alternatives to Base/SAS
Content-Type: text/plain; charset=us-ascii
Walt Smith wrote:
[okay, this is me, unattributed]
> > Definitely. For Perl one uses the modules. So instead of
> > writing the six lines to get the quartiles out of proc univariate
> > and into another dataset, one uses two lines to get the
> > quartiles out of the Statistics::Descriptive module and into
> > an array.
> I could be wrong - I'm just learning Perl, but have been doing SAS for
> 16 years - but I'd challenge any Perler to do the equivalent of:
> proc arima data=datalib.mydata;
> identify var=y(1)(12) crosscor=(x1(1)(12));
> estimate p=1 q=1 input=(2$(1,3)x1);
> forecast id=date interval=month out=fcstdata lead=36;
Not in 5 lines, but one can do this with modules like PDL
[which is more like PROC IML than PROC ARIMA actually].
But I prefer SAS for canned procedures. OTOH, SAS is a lot
less helpful than Perl for something like the sample
variance of a non-trivial probability-based sample design,
for example a spatially-hierarchical randomization (Don
Stevens, 1998). And variances are one of the areas in which
we expect SAS to excel.
> proc gplot data=fcstdata;
> plot forecast*date y*date / overlay frame grid;
If you want superior graphical capabilities, look at some of
the available Perl modules, like:
But graphics is not one of the strong points of SAS, other
than what can be generated by the PROCs.
> And thats just the simple stuff!!
Of course, these are the strengths of SAS, and supposedly
the weaknesses of general programming languages like Perl
[or Java, or Python, or whatever]. You seem to have missed
the point of my discussion.
David Cassell, OAO email@example.com
Senior computing specialist