How to use xtabs in a user defined function?

writer_typer 0
Viewed 36

I'm trying to create a wrapper for xtabs with a custom function.

I tried doing this:

xtabs_fun <- function(data, variable) {
  xtabs(~ variable, data = data)
}

But using it doesn't return the correct value.

xtabs_fun(mtcars, "cyl")

cyl
1

The expected output should be this:

xtabs(~ cyl, data = mtcars)

cyl
 4   6   8
11   7  14
r
3 Answers

We could use reformulate

xtabs_fun <- function(data, variable) {
    xtabs(reformulate(variable), data = data)
}

-testing

> xtabs_fun(mtcars, "cyl")
cyl
 4  6  8 
11  7 14
akrun0

xtabs accepts a data frame so:

xtabs_fun <- function(data, variable) xtabs(data[variable])

# test
xtabs_fun(mtcars, "cyl")
## cyl
##  4  6  8 
## 11  7 14

Alternately use table which works the same:

table_fun <- function(data, variable) table(data[variable])
G. Grothendieck0

Variable names inside formulas are always taken literally rather than being substituted, so inside your function, xtabs is looking for a column called variable inside mtcars, which doesn't exist. Instead, you can build the formula as a string:

xtabs_fun <- function(data, variable) {
  xtabs(as.formula(paste("~", variable)), data = data)
}

xtabs_fun(mtcars, "cyl")
#> cyl
#>  4  6  8 
#> 11  7 14

If you prefer your function to take unquoted variable names in a tidyverse style, you can do

xtabs_fun <- function(data, variable) {
  xtabs(as.formula(paste("~", deparse(substitute(variable)))), data = data)
}

mtcars |> xtabs_fun(cyl)
#> cyl
#>  4  6  8 
#> 11  7 14

Created on 2022-09-24 with reprex v2.0.2

Allan Cameron0
Related