Apply Function Using Multiple Changing Arguments in R

Question

I want to apply a function over all rows referencing multiple columns in a data frame in R. So, for example, if I have a data frame called "data" with three variables "var1", "var2", and "var3" and I want to apply a function to each row:

myfunc <- function(var1, var2, var3)){
result <- var1*var2*var3
return(result)
}

Then the pseudocode would be:

apply(data, myfunc(data$var1, data$var2, data$var3))

This code does not work, however, because data is a data frame, not a vector, and lapply does not seem able to take more than one vector. How do I make this work? I am open to any type of solution, but I have to be able to reference multiple changing arguments and call a predefined function.

Solution 1

No need for apply:

#define function
myfunc <- function(var1, var2, var3){
  result <- var1*var2*var3
  return(result)
}
#dummy data
dat <- data.frame(A=c(1,2,3),B=c(4,5,6),C=c(7,8,9))

#test function
myfunc(dat$A,dat$B,dat$C)

#output
[1]  28  80 162

Solution 2

You seem to be fairly close, but you're missing a little bit in the call.

apply(data, 1, function(x,y,z) myfunc(data$var1,data$var2,data$var3))

seems to do the trick.

edit: I tested this on the data set

data<-data.frame(a=c(1,2,3),b=c(4,5,6),c=c(7,8,9))

and got the output

[,1] [,2] [,3]
[1,]   28   28   28
[2,]   80   80   80
[3,]  162  162  162

Solution 3

Here is an untested (no reproducible example provided) suggestion:

myfunc <- function(x)){
    result <- x$var1*x$var2*x$var3
    return(result)
}
apply(data, 1, myfunc)

Comments

[QuantIbex]

As pointed out by @zx8754 in his answer, you might avoid the use of apply. But this will be possible only if the computation in the function is vectorized, which is the case in the example you provided. Is it the case with the computation you want to do (myfunc seems to be for illustration purpose only)?