R help - Sep 1998 - R-beta: Applying functions to vectors--question

Here is a problem I have encountered with R when calling functions
with vector arguments,
when the vectors contain some values which may be nonsensical as
arguments.

Rather than apply the function piecewise to each element of the
vector, it does something
more complicated (and less helpful).

I have constructed a function f (below) which tests if its argument is
negative; if it is,
it returns 0, if not, it returns sqrt(f).

This works fine with scalar arguments.

If you give it a vector argument, however, it seems to have trouble. 
What appears to 
happen is that the whole vector is evaluated in the if statement and
in the then part.

If the first value in the vector is a nonsensical argument (negative),
 the if is not true, so the function returns a single 0.

If the first value in the vector is a good argument (positive), then
it returns the
sqrt function, applied to each element of the input vector.

So what I'm wondering is -- how can you simply apply the function f to
each element 
of the vector without this weird thing happening?

Not having used S, I don't know whether this is 

(a) a design feature that is there for a good reason that I just
haven't figured out
(b) a bug in S that has been reproduced in R
or
(c) a bug unique to R.


Can someone help me out?  Many thanks in advance

Marc Lipsitch
lipsitch at biology.emory.edu
> f_function(e) {
+ 
+ out_0
+ 
+ if(e>0)
+ 
+          out_sqrt(e) 
+ 
+ out}
> 
> 
> 
> f(-2)
[1] 0
> 
> f(-2:2)
[1] 0
> 
> f(2:-2)
Warning: NaNs produced in function "sqrt"
[1] 1.414214 1.000000 0.000000      NaN      NaN
> f
function (e) 
{
	    out <- 0
	    if (e > 0) 
	    	    out <- sqrt(e)
	    out
}

_________________________________________________________
DO YOU YAHOO!?
Get your free @yahoo.com address at http://mail.yahoo.com

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Marc Lipsitch wrote:
> I have constructed a function f (below) which tests if its argument is
> negative; if it is,
> it returns 0, if not, it returns sqrt(f).
> 
> This works fine with scalar arguments.
> 
> If you give it a vector argument, however, it seems to have trouble.
You want to use the "ifelse" function,  to apply a condition
element-wise
to a vector, e.g.

f <- function(e)
	ifelse(e>0,sqrt(e),0)

Or, to avoid the warning (which comes about because both alternative
answers
are evaluated before deciding which one to apply)

f <- function(e)
	sqrt(ifelse(e>0,e,0))

NB The help page for "if" points to the help page for
"ifelse" which
gives
an example very similar to this.

Martyn
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at
stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._