[R] reading Systat into R

Jonathan Baron baron at psych.upenn.edu
Fri Oct 8 21:13:00 CEST 2004


On 10/08/04 12:02, Jacob Wegelin wrote:
>
>How do I read a Systat file into R?  The  following email by Marc Schwartz
>http://tolstoy.newcastle.edu.au/R/help/04/06/1005.html deals with reading Systat on
>a Linux machine.  I'm running R on Windows (precise version info below).  A colleague
>sent me data in systat, a *.SYD file.  Do I have to ask the colleague to re-save the
>data as Excel or text?

Unfortunately, yes, to my knowledge.  Text, of course, could be
SYC, or one of the various outputs that Systat produces.

Because I used to use Systat, and many of my colleagues still do,
I spent some time investigating this problem.  Here are some
excerpts from correspondence.  Because one was just to me and
not the list, I'm removing the From.  I have the code, but I
could not get it to work.  I put it aside.  I think it needs
someone who is more familiar with c than I am (which doesn't take
much).

Jon
---

>>Contact David.Baird at AgResearch.CO.NZ because his dataload program will
>>import from SYSTAT and output to an .rda file.  But I fear that dataload
>>is no longer available to new users, only those grandfathered in before
>>a license issue came up.  -Frank
>
> Thanks.  DataLoad does have a free version, but the whole thing
> is Windows only.  (Just search for dataload on Google.)  If I had
> Windows, I'd have Systat.

No, David gave me the ready-to-run linux executable.  Never used it on 
windows.

--

I've put three header files and two pre-ANSI C files in a tarball. I note
that sys_errlist causes an error under compilation now, but for code last
looked at 12 years ago, that's not bad. The reading code was written by me
alone based on reverse engineering, but when I had a question - as you'll
see from the code - someone I think called P. Fleury from Systat sent me
some C that they'd been trying out, which I saw after writing my own. I
think mine was more thorough. 

The endian-ness was because I needed to read and write on non-Intel
machines (you may need the functions I used then - please ask if so, or
substitute more modern ones if required). Please treat this as GPL - I
think there was not difficulty with Systat at the time, and they knew that
I had this on my ftp server then (a 1992 paper in Computers and
Geosciences).  I still have *.sys files that I could try it on if it works
- please let me know if there are unresolved calls (walert() wrote an
error message, for example).

Best wishes,

Roger [Bivand]

>From r-help-bounces at stat.math.ethz.ch  Wed Jun 16 11:33:01 2004

The commercial package dbmscopy has a Linux version. 
I have used dbmscopy for several years and have been happy 
with it as it converts data files among many spreadsheets and 
statistics programs.

http://www.conceptual.com/dbmscopt.htm

However, somewhat recently they were purchased by SAS, so 
I'm not sure of current state of the program. There are 
probably other commercial packages as well.

Anne

Hi Jon and Anne!

One other commercial product to check out is Stat/Transfer.  More
information on supported formats is at:
http://www.stattransfer.com/html/formats.html

They do support Windows, MacOS and Unix/Linux. Demo downloads are
available from: http://www.stattransfer.com/html/download.html

Unix/Linux pricing is available at:
http://www.stattransfer.com/html/prices_-_unix.html.

HTH,

Marc Schwartz


-- 
Jonathan Baron, Professor of Psychology, University of Pennsylvania
Home page: http://www.sas.upenn.edu/~baron




More information about the R-help mailing list