dplyr mutating multiple columns using two columns as arguments to custom function

Question

I have a data frame:

df <- data.frame(A=c(10, 20, 30, 20), 
      B=c(0, 10, 20, 10), 
      C=c(11, 22, 21, 12),
      D=c(13, 11, 33, 15))

 A  B  C  D
10  0 11 13
20 10 22 11
30 20 21 33
20 10 12 15

and a function to get the index of the number in a pair closest to a number of interest:

comp <- function(x, y) {
    # x = number of interest, y = vector of 2 numbers for comparison)
    ind <- which(abs(y-x)==min(abs(y-x)))
    if (length(ind)==2) {
        ind <- 3
    }
    return(ind)
}

(The if statement is for when the number is smack in the middle of the two numbers, ex. 15 compared to 10 and 20).

I would like to change columns C & D to the index for which the number is closest using my function (1 for A or 2 for B):

 A  B  C  D
10  0  1  1
20 10  1  2
30 20  2  1
20 10  2  3

I'm not sure how to call in columns A and B as arguments for the function. I've tried mutate_at(df, 3:4, funs(comp), c(df$A, df$B)), but that returns:

Doesn't have to be a tidyr solution, whatever works! Thanks

Your function only takes two arguments. What should x and y be? A and B, or A & C and B & C, etc? — acylam
– acylam, Commented Jul 5, 2018 at 19:05
x is the number of interest, y is a vector of 2 numbers. Ex. comp(13, c(10, 20)) returns 1, since that is the index of the closest number in the pair. So for the df, c(df$A, df$B) is my attempt at calling the y argument, since the x argument should be an individual column (C or D). — hmg
– hmg, Commented Jul 5, 2018 at 19:07

zack · Accepted Answer · 2018-07-05 19:12:51Z

2

I changed around your function a little bit in order for vectorization to work. It also only accepted 2 values, when you were looking to compare 1 value with 2 others, so 3 arguments would be needed:

comp <- function(val, x, y){
  case_when(
    abs(val - x) < abs(val - y) ~ 1,
    abs(val - x) > abs(val - y) ~ 2,
    TRUE ~ 3)
}

df %>% 
  mutate_at(vars(C,D), comp , .$A, .$B)

   A  B C D
1 10  0 1 1
2 20 10 1 2
3 30 20 2 1
4 20 10 2 3

answered Jul 5, 2018 at 19:12

zack

5,4352 gold badges23 silver badges26 bronze badges

Sign up to request clarification or add additional context in comments.

1 Comment

hmg Over a year ago

Perfect, thanks for the introduction to case_when!

Collectives™ on Stack Overflow

dplyr mutating multiple columns using two columns as arguments to custom function

1 Answer 1

1 Comment

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Related