r - passing formula in a list to `rlang::exec`

Question

I am trying to use rlang::exec in custom functions where I want to pass additional arguments as a list and then splice them. Usually this works without any problem. But I am encountering problems while doing this routine when there is a formula argument involved.

without list

library(rlang)

exec(
  .fn = stats::t.test,
  formula = wt ~ am,
  data = mtcars
)
#> 
#>  Welch Two Sample t-test
#> 
#> data:  wt by am
#> t = 5.4939, df = 29.234, p-value = 6.272e-06
#> alternative hypothesis: true difference in means between group 0 and group 1 is not equal to 0
#> 95 percent confidence interval:
#>  0.8525632 1.8632262
#> sample estimates:
#> mean in group 0 mean in group 1 
#>        3.768895        2.411000

with list

extra.args <- list(formula = wt ~ am) 

exec(
  .fn = stats::t.test,
  data = mtcars,
  !!!extra.args
)
#> Error in t.test.default(data = structure(list(mpg = c(21, 21, 22.8, 21.4, : argument "x" is missing, with no default

How can I get this to work?

Answer 1

I'm not sure this is rlang::exec's fault. The problem really has to do with S3 dispatch and the fact that different functions are called based on the class of the first parameter, not the names of the parameters. With your current calling method, you are passing data= before your formula. This also causes a problem when calling the function directly

stats::t.test(data=mtcars, formula=wt~am)

The easiest way to get around this would be to pass the parameters in the "natural" order for proper S3 dispatch to take place

extra.args <- list(formula = wt ~ am) 
exec(
  .fn = stats::t.test,
  !!!extra.args,
  data = mtcars
)

or leave the formula parameter unnamed so it becomes the first unnamed-parameter.

extra.args <- list(wt ~ am) 

exec(
  .fn = stats::t.test,
  data = mtcars,
  !!!extra.args
)