mapply with multiple arguments where one argument is constant (data)

Question

I'm struggling with using mapply on functions I construct where I have one or more arguments that are needed because I am programming in a bigger environment, for example if I write a function where one of the arguments are data.

fun_test <- function(data,col,val1,val2){return(data[col][1,] * val1-val2)}

So data and col can for example be constant, but I want to vary the output of my function depending on val1 and val2:

> mapply(FUN=fun_test,mtcars,"cyl",mtcars$cyl,mtcars$cyl*2)
Error in data[col][1, ] : incorrect number of dimensions

I'm trying to understand how mapply works; I surely cannot pass mtcars, and "cyl" as a vector, can I?

EDIT: I have an environment in which the data may vary, e.g. sometimes I use mtcars, sometimes it is another dataset. So I cannot hardcode the data into the function

EDIT2: 1) I have data some dataset, 2) I have different Excel-files that I read into R, 3) I make a lookup function that extracts information from these Excel-files in R, 4) for one or two variables (from the dataset) at the time I go into the lookup-functions I created and extract information.

So these lookup functions depend on both the data (the variables I need to lookup) and the Excel-files that I use to do the looking up.

Answer 1

mapply is a multidimensional lapply. This means that instead of iterating over just one object (i.e. the columns of a data.frame or the elements of a vector), it iterates over multiple ones at the same time. The only condition is that the length of those objects needs to be the same, i.e. the columns of a data.frame and the lengths of the vectors. So, you cannot pass constants (unless you pass in a vector of the same constants to match the length, but why would you do that).

Try an easy example (sums the same indexes of the vectors):

mapply(sum, 1:10, 11:20)

So, in your case, just pass in the constants straight into the function:

fun_test <- function(val1, val2){return(mtcars['cyl'] * val1 - val2)}

mapply(FUN=fun_test, mtcars$cyl, mtcars$cyl*2)

Update:

Then I think what you need is to include mapply within your function. In that way you can add any argument you like (both constants and variable). It would look like this:

myfunc <- function(data, col, val1, val2) {

  fun_test <- function(val1, val2) {
    data[col] * val1 - val2 
  }

  mapply(FUN=fun_test, val1, val2)

}

myfunc(mtcars, 'cyl', mtcars$cyl, mtcars$cyl*2)