[R] SAS or R software

Frank E Harrell Jr f.harrell at vanderbilt.edu
Sat Nov 20 23:05:30 CET 2004

BXC (Bendix Carstensen) wrote:
> Two major advantages of SAS that seems to have been overlooked in
> the previous replies are:
> 1) The data-set language is SAS for data manipulation is more
>    human-readable than R-code in general. 
>    R is not a definite write-only laguage as APL, but in particular
>    in datamanipulation it is easy to write code that is impossible
>    to decipher after few weeks. 
>    You can also produce unreadable code in SAS, but it generally takes 
>    more of an effort.
>    Thus: Data manupulation is easier to document in SAS than in R.

I agree with the readability of data manipulation code, especially for 
novice users.  As far as functionality for data manipulation is 
concerned, R has more flexibility and is faster to program once one has 
experience.  I do think that learning data manipulation techniques in R 
takes a while.

> 2) proc tabulate.
>    This procedure enables you to do extensive sensible tabulation
>    of your data if you are prepared to read the manual.
>    (This is not a product of the complexity of the software,
>     but of the complexity of the tabulation features).
>    Compared to this only rudimentay tools exist in R (afaik).

I could not disagree more with that statement.  Look for example at 
http://biostat.mc.vanderbilt.edu/twiki/pub/Main/StatReport/summary.pdf . 
  With R you can customize tables by specifying your own function for 
computing any statistic of interest.  Let's see how to use PROC TABULATE 
to display stratified Kaplan-Meier 4-year survival estimates along with 
median and median life length, not to mention match the fine control of 
formatting available using the combination of R and LaTeX.

> So if you want to do well documented data manipulation and clear
> and compact tables go for SAS.

Readability is different from being well-documented.  And for clear and 
compact tables, R is the winner hands down.

Frank Harrell

> If you want to do statistical analyses and graphics (in finite time)
> go for R.
> Bendix Carstensen
> ----------------------
> Bendix Carstensen
> Senior Statistician

Frank E Harrell Jr   Professor and Chair           School of Medicine
                      Department of Biostatistics   Vanderbilt University

More information about the R-help mailing list