[R] Help with error in "if then" statement

Erik Iverson iverson at biostat.wisc.edu
Fri Jul 11 22:30:50 CEST 2008

Hello -

Andrew Rominger wrote:
> Dear list,
> I'm afraid this is a mundane question.  Here's the background: I've 
> produced a function which allows me to sequentially measure angles and 
> distances from a specified reference point to a curve defined by 
> empirical data points while (i.e. using a while loop) the angle being 
> measured is within certain bounds.  Because the curve is circular I need 
> to pars the data into basically an "upper" curve and "lower" curve.  I 
> tried to do this with an if statement, specifically:
> ycrit<-subset(data,subset=data$x==min(data$x)

You might consider not using 'data' as a variable name (it is the name 
of a function in the utils package) , and clearer code might be

ycrit <- subset(data, x == min(x))

But what does this return?  It can definitely return a data.frame as in 
the following example:

t1 <- data.frame(a = rep(2,2))
class(subset(t1, a == min(a)))

You probably do not want to them subsequently use this returned subset 
in an 'if' clause.  Also, consider cases where there are NAs in the data:

t2 <- data.frame(a = c(NA,rep(2,2)))
subset(t2, a == min(a))

This code returns a data.frame with 0 rows.

I don't know what your specific problem is, but hopefully this might 
lead you in the right direction.

> y.max<-length*sin(angle)+y.ref #length, angle and y.ref are given
> if(y.max<ycrit){
> ........
> }
> Now the problem:  The while loop works for 4 iterations, until I get the 
> error message:
> "Error in if (y.max < ycrit) { : missing value where TRUE/FALSE needed"
> When I forced the function to print y.max and ycrit for each iteration 
> of the while loop, it returns finite real numbers, including ycrit = 
> 153.5 and y.max = 245.16 for the step which returns the error message.
> Any ideas about what's going on here--why does R "think" that 
> 245.16<153.5 is "missing," or is anything other than TRUE/FALSE?  Am I 
> using "if" incorrectly?  In which case would it be more appropriate to 
> perhaps create subsets of the data points based on < or > ycrit?
> Thanks in advance for any guidance--
> Andy

More information about the R-help mailing list