Apply variable function to columns in data.table

Question

I'm wondering if there's a way to apply a function in a string variable to .SD cols in a data.table.

I can generalize all other parts of function calls using a data.table, including input and output columns, which I'm very happy about. But the final piece seems to be applying a variable function to a data.table, which is something I believe I've done before with dplyr and do.call.

mtcars <- as.data.table(mtcars)

returnNames <- "calculatedColumn"

SDnames <- c("mpg","hp")

myfunc <- function(data) {
    print(data)
    return(data[,1]*data[,2])
}

This obviously works:

mtcars[,eval(returnNames) := myfunc(.SD),.SDcols = SDnames,by = cyl]

But if I want to apply a dynamic function, something like this does not work:

functionCall <- "myfunc"

mtcars[,eval(returnNames) := lapply(.SD,eval(functionCall)),.SDcols = SDnames,by = cyl]

I get this error:

Error in `[.data.table`(mtcars, , `:=`(eval(returnNames), lapply(.SD,  :  attempt to apply non-function

Is using "apply" with "eval" the right idea, or am I on the wrong track entirely?

This is potentially helpful...stackoverflow.com/questions/29620783/… But something like this doesn't work: mtcars[,eval(returnNames) := lapply(.SD,function(u) {sapply(functionCall, function(f) do.call(f,list(u)))}),.SDcols = SDnames,by = cyl] — shwan
– shwan, Commented Oct 30, 2018 at 14:17

IceCreamToucan · Accepted Answer · 2018-10-30 14:54:45Z

3

You don't want lapply. Since myfunc takes a data.table with multiple columns, you just want to feed such a data table into the function as one object.
To get the function you need get instead of eval
On the left-hand-side of :=, you can just put the character vector in parentheses, eval isn't needed

-

mtcars[, (returnNames) := get(functionCall)(.SD)
       , .SDcols = SDnames
       , by = cyl]

head(mtcars)
#     mpg cyl disp  hp drat    wt  qsec vs am gear carb calculatedColumn
# 1: 21.0   6  160 110 3.90 2.620 16.46  0  1    4    4           2310.0
# 2: 21.0   6  160 110 3.90 2.875 17.02  0  1    4    4           2310.0
# 3: 22.8   4  108  93 3.85 2.320 18.61  1  1    4    1           2120.4
# 4: 21.4   6  258 110 3.08 3.215 19.44  1  0    3    1           2354.0
# 5: 18.7   8  360 175 3.15 3.440 17.02  0  0    3    2           3272.5
# 6: 18.1   6  225 105 2.76 3.460 20.22  1  0    3    1           1900.5

The code above was run after the following code

mtcars <- as.data.table(mtcars)

returnNames <- "calculatedColumn"

SDnames <- c("mpg","hp")

myfunc <- function(data) {
    print(data)
    return(data[,1]*data[,2])
}

functionCall <- "myfunc"

answered Oct 30, 2018 at 14:54

IceCreamToucan

28.8k2 gold badges27 silver badges48 bronze badges

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

1 Comment

shwan Over a year ago

That's pretty elegant @RyanD...thanks! Hadn't seen a get used before on a data.table

Collectives™ on Stack Overflow

Apply variable function to columns in data.table

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