double if conditioning in r language syntax

Question

What is the correct way of writing if syntax on R?

With:

for ( i in 1:200 ) {
     if( (time[i]<731) | (time[i]==NA) ) { x[i] <- NA }
}

I keep getting:

missing value where TRUE/FALSE needed

The idea is to cut the beginning of the time series to calculate statistics. x and time are both numeric. Also this question is not an exact duplicate, the other post was dealing with a && statement and no foo==NA but foo==0. I thought that | or || would work like (TRUE | NA) = TRUE but it seems I was wrong.

score 2 · Accepted Answer · answered Dec 16 '13 at 15:44

2

you can't write things like if (foo==NA) because comparing anything with NA returns NA. Instead, try

if( is.na(foo))

answered Dec 16 '13 at 15:44

Carl Witthoft

20,036
8
41
71

@joran very gracious to delete your response -I thought you beat me to it. – Carl Witthoft Dec 16 '13 at 15:45
Yours was already there when I hit submit. Joshua found a good duplicate anyway. – joran Dec 16 '13 at 15:48
if( (time[i]<731) | is.na(time[i]) ) { x[i] – user3083324 Dec 16 '13 at 20:03

Sven Hohenstein · Answer 2 · 2013-12-16T17:18:52.243

0

Try

is.na(x) <- (time < 731 | is.na(time))

edited Dec 16 '13 at 17:18

answered Dec 16 '13 at 15:54

Sven Hohenstein

78,180
16
134
160

this doesn't work, all it does is make x full of NA from beginning to end. – user3083324 Dec 16 '13 at 17:12
@user3083324 There was a typo. Try the updated version. – Sven Hohenstein Dec 16 '13 at 17:19

score 0 · Answer 3 · answered Dec 16 '13 at 16:01

Assuming you wanted:

> n = 10
> x <- rep(0, n)
> time <- rpois(n, 731)
> 
> for ( i in 1:n ) {
+   if( time[i] < 731 | is.na(time[i])) {  
+     x[i] <- NA
+   } 
+ }
> x
 [1]  0  0 NA NA NA  0 NA NA  0  0

You can do:

> ifelse(time < 731 | is.na(time[i]),  NA, x)
 [1]  0  0 NA NA NA  0 NA NA  0  0

double if conditioning in r language syntax

3 Answers3

Linked