generic handling of NA and NaN and NULL

The version I gave is obviously not vectorized (since your original version
seem to indicate that the argument won't have length > 1, otherwise the if()
won't really make sense).

Replacing is.null(n) with length(n)!=1 (or length(n)==0) should do the trick
(I hope!).

Andy

-----Original Message-----
From: Robin Hankin [mailto:r.hankin at auckland.ac.nz]
Sent: Thursday, February 13, 2003 6:30 PM
To: andy_liaw at merck.com
Cc: r.hankin at auckland.ac.nz; r-help at stat.math.ethz.ch
Subject: Re: [R] generic handling of NA and NaN and NULL




Hello Andy

thanks for this; but


R> x <- 1:10
R> f
function(n) {
 if(is.null(n) || is.na(n) || abs(n) < pi) {
   return(FALSE)
 } else {
   return(TRUE)
 }
}
R> x <- 1:10
R> f(x[x>11])
Error in if (is.null(n) || is.na(n) || abs(n) < pi) { : 
missing value where logical needed

Try:

f <- function(n) {
 if(is.null(n) || is.na(n) || abs(n) < pi) {
   return(FALSE)
 } else {
   return(TRUE)
 }
}

Note that the order of the conditions inside if() matters: 

is.na(n) only

gets evaluated if is.null(n) is FALSE, and so on.

Andy

-- 

Robin Hankin, Lecturer,
School of Geography and Environmental Science
Tamaki Campus
Private Bag 92019 Auckland
New Zealand

r.hankin at auckland.ac.nz
tel 0064-9-373-7599 x6820; FAX 0064-9-373-7042

------------------------------------------------------------------------------