function to set NA's in a specific column of a data frame to 0 in R

Question

I am trying to write a function to replace the NA values in the second column of a data frame to 0. I know if "x" is my data frame then

x[,2][is.na(x[,2]) <- 0 

will do the job. However, I wanted to write a function for this so I can apply to different data frames. I came up with

SetNaToZero <- function(x) {
x <-data.frame(x)
x[1:nrow(x),2][is.na(x[1:nrow(x),2])] <-0
return(x)
}

But it returns the following error when I apply it on a data frame:

Error in `[<-.data.frame`(`*tmp*`, 1:nrow(x), 2, value = numeric(0)) : 
replacement has length zero 

I appreciate any suggestions.

Answer 1

I think you are making it harder than you need to. The code you have in the first chunk there would be fine as a function:

SetNaToZero <- function(x) {
  x[,2][is.na(x[, 2])] <- 0
  return(x)
}

In action:

set.seed(123)
dat <- data.frame(a=rnorm(10),
                  b=sample(c(NA, 1:3), 10, replace=T))

SetNaToZero(dat)
             a b
1  -0.56047565 3
2  -0.23017749 2
3   1.55870831 2
4   0.07050839 3
5   0.12928774 2
6   1.71506499 2
7   0.46091621 2
8  -1.26506123 2
9  -0.68685285 1
10 -0.44566197 0

And if you want to modify the object, you can always do dat <- SetNaToZero(dat)

Answer 2

You can simplify the function:

SetNaToZero <- function(x) {
  x[is.na(x[[2]]), 2] <- 0
  x
}

Simplify it further:

SetNaToZero <- function(x)
  "[<-"(x, is.na(x[[2]]), 2, 0)