A function that returns TRUE on NA, NULL, NaN in R?

Question

Is there a function in R that behaves like this:

isnothing = function(x) {
    is.null(x) | is.na(x) | is.nan(x)
}

Answer 1

I was also missing such a function and added this to my .Rprofile ages ago. If someone knows of a base function that does the same thing I also want to know.

is.blank <- function(x, false.triggers=FALSE){
    if(is.function(x)) return(FALSE) # Some of the tests below trigger
                                     # warnings when used on functions
    return(
        is.null(x) ||                # Actually this line is unnecessary since
        length(x) == 0 ||            # length(NULL) = 0, but I like to be clear
        all(is.na(x)) ||
        all(x=="") ||
        (false.triggers && all(!x))
    )
}

As @shadow mentioned, NA, NaN and NULL have different meanings that are important to understand. However, I find this function useful when I make functions containing optional arguments with default values, that I want to allow the user to suppress by setting them to any "undefined" value.

One such example is xlab of plot. I can never remember if it is xlab=NA, xlab=FALSE, xlab=NULL or xlab="". Some produce the desired result and some don't, so I found it convenient to catch all with the above function when I develop code, particularly if other people will use it too.

Answer 2

I believe you are basically looking for what gtools invalid() function does.

 ?gtools::invalid

For example,

gtools::invalid(NA)
[1] TRUE
gtools::invalid(NULL)
[1] TRUE
gtools::invalid(NaN)
[1] TRUE

Answer 3

I had a very similar problem too. I was trying to check whether row names of a given matrix or data frame are proper.

x <- matrix(1:6,ncol=2)
isnothing = function(x) {
  any(is.null(x))  | any(is.na(x))  | any(is.nan(x)) 
}
isnothing(rownames(x))

The function will throw an error. But when I used short circuiting and changed it to:

isnothing = function(x) {
  any(is.null(x))  || any(is.na(x))  || any(is.nan(x)) 
}
isnothing(rownames(x))

It solved my problem. I guess checking first nullity and then going forward to check other cases if it is FALSE solved my issue. I checked it with couple problematic cases that I can think of and it worked. I don't know whether there are exceptions to it but it worked for my purposes for now.

rownames(x) <- c("a",NaN,NA)
isnothing(rownames(x))