Function in data.table with two columns as arguments

Question

I have the following function:

DT <- data.table(col1 = 1:4, col2 = c(2:5))

fun <- function(DT, fct){
  DT_out <- DT[,new_col := fct]
  return(DT_out)
}

fun(input, fct = function(x = col1, y = col2){y - x})

In reality I have some processing before and after this code snippet, thus I do not wish to use directly the statement DT[,new_col := fct] with a fixed fct (because the fct should be flexible). I know this question is very similar to this one, but I cannot figure out how to reformulate the code such that two columns as arguments for the function are allowed. The code above gives the error:

Error in `[.data.table`(DT, , `:=`(new_col, fct)) : 
  RHS of assignment is not NULL, not an an atomic vector (see ?is.atomic) and not a list column.

Do you want new_col := fct(col1,col2)? As it is, you're assigning a function to new_col, not it's output — Rohit
– Rohit, Commented Mar 20, 2020 at 16:28
No that won't satisfy my needs, because I want the arguments to be flexible like in the answer given below! But thanks anyway — yaennu.s
– yaennu.s, Commented Mar 20, 2020 at 18:54

IceCreamToucan · Accepted Answer · 2020-03-20 16:42:44Z

3

One option if you don't mind adding quotes around the variable names

fun <- function(DT, fun, ...){
  fun_args <- c(...)
  DT[,new_col := do.call(fun, setNames(mget(fun_args), names(fun_args)))]
}

fun(DT, fun = function(x, y){y - x}, x = 'col1', y = 'col2')

DT
#    col1 col2 new_col
# 1:    1    2       1
# 2:    2    3       1
# 3:    3    4       1
# 4:    4    5       1

Or use .SDcols (same result as above)

fun <- function(DT, fun, ...){
  fun_args <- c(...)
  DT[, new_col := do.call(fun, setNames(.SD, names(fun_args))), 
     .SDcols = fun_args]
}

edited Mar 20, 2020 at 16:42

answered Mar 20, 2020 at 16:37

IceCreamToucan

28.8k2 gold badges27 silver badges48 bronze badges

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

1 Comment

yaennu.s Over a year ago

Yesss! Both are nice solutions, but I like the second one with .SDcols better. It is clearer and more in the style of data.table in my opinion. And thanks a lot for pointing out the three-dots construct, this opens a whole new world for me!

Collectives™ on Stack Overflow

Function in data.table with two columns as arguments

1 Answer 1

1 Comment

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

1 Comment

Your Answer

Sign up or log in

Post as a guest

Linked

Related