Date: Mon, 5 Mar 2012 12:50:23 +0000
Reply-To: "Burgess, Otto" <OBurgess@ATPCO.NET>
Sender: "SAS(r) Discussion" <SAS-L@LISTSERV.UGA.EDU>
From: "Burgess, Otto" <OBurgess@ATPCO.NET>
Subject: Re: base sas vs. r
Content-Type: text/plain; charset="iso-8859-1"
Actually, yes you can and it would be my choice based on what I have read
Anyone out there using WPS on z/os care to share anything about use and transition?
From: Søren Lassen [mailto:s.lassen@POST.TELE.DK]
Sent: Monday, March 05, 2012 4:04 AM
To: SAS-L@LISTSERV.UGA.EDU; Burgess, Otto
Subject: Re: base sas vs. r
You can get WPS for the mainframe as well, why is it not an option?
On Fri, 2 Mar 2012 12:58:09 +0000, Burgess, Otto <OBurgess@ATPCO.NET> wrote:
>I think I should have used a different subject line
>I should have said mainframe, z/os SAS/base vs. r
>WPS is not a mainframe option
>However, I do like the fact the code would not need to change with WPS
>From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
>Sent: Friday, March 02, 2012 1:46 AM
>Subject: Re: base sas vs. r
>Yes there are alternative ways to do things. Instead of coding SAS you
do a lot of stuff in other databases, stat packages etc.
>However if you are familiar with SAS programming it will take some time
before you are productive in a new language.
>R is great for statistics but not so good for data management.
>If your goal is to save license money then you should take a look at WPS.
WPS is an interpreter that will run most of your existing SAS code without any changes.
>But the license cost is only a fraction of SAS.
>Den 01/03/2012 kl. 21.02 skrev Joe Matise:
>> If you're only using SAS/BASE, then you might be able to just use
>> MySQL or SQLLite or something as an alternative (if you're doing more
>> data manipulation). But again it of course depends on the details.
>> On Thu, Mar 1, 2012 at 1:38 PM, Burgess, Otto <OBurgess@atpco.net> wrote:
>>> Actually, that is helpful information
>>> I don't think much data is being processed but I am not sure
>>> If a SAS program is processing a huge dump for something, that could
>>> be a serious problem
>>> I am also seeing that the syntax is different enough to require
>>> - a serious drawback
>>> I know they are not doing serious statistical analysis, since we
>>> only have SAS/Base
>>> -----Original Message-----
>>> From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf Of
>>> Nordlund, Dan (DSHS/RDA)
>>> Sent: Thursday, March 01, 2012 2:29 PM
>>> To: SAS-L@LISTSERV.UGA.EDU
>>> Subject: Re: base sas vs. r
>>>> -----Original Message-----
>>>> From: SAS(r) Discussion [mailto:SAS-L@LISTSERV.UGA.EDU] On Behalf
>>>> Of Burgess, Otto
>>>> Sent: Thursday, March 01, 2012 11:04 AM
>>>> To: SAS-L@LISTSERV.UGA.EDU
>>>> Subject: base sas vs. r
>>>> So we have base SAS and only base SAS licensed on one LPAR
>>>> We are debating moving some test systems from this LPAR to another
>>>> and since they use SAS, it would need to be moved too, along with
>>>> any processes that use SAS
>>>> I have suggested looking at using R instead of SAS
>>>> Any thoughts?
>>>> I have never used R myself
>>>> Does R mostly cover 100% of at least base SAS?
>>> Well, the short answer to your question about coverage is yes (...
>>> uhh, maybe?).
>>> R is a very capable programming platform, albeit very different in
>>> philosophy from SAS. Pretty much everything you might like to do in
>>> SAS can be done in R, some things more easily, some more difficult.
>>> The major limitation in R comes from the fact that R processes
>>> everything in memory. So as a practical matter you are limited to
>>> the amount of memory you have available. In addition, 32-bit
>>> integers are used for indexing objects (whether you are working on
bit or 64-bit platforms).
>>> There are some projects working on high performance computing and
>>> big data both in the commercial and non-commercial sectors, but
>>> haven't followed them closely.
>>> Without knowing a whole lot more about what resources you have in
>>> terms of hardware and programmers, and legacy code, and what size
>>> data you work with, and ..., it is hard to know whether R would be
>>> viable for your purposes.
>>> Sorry, I can't be of much more help,
>>> Daniel J. Nordlund
>>> Washington State Department of Social and Health Services Planning,
>>> Performance, and Accountability Research and Data Analysis Division
>>> Olympia, WA 98504-5204