如何编写也适用于 dplyr::across 的自定义函数？

Question

我编写了一些自定义函数（请参阅下面的示例），这些函数在使用基本 R 和 dplyr::mutate 管道时都可以工作。然而，我很难让它与 dplyr::across 一起动态工作。

我做错了什么？

test1 <- function(data, column)
{
  if (!column %in% names(data))
  {
    stop("error in test1")
  }
  return(TRUE)
}

test2 <- function(data, column)
{
  check <- test1(data, column)
  
  if (check == TRUE)
  {
    my_sum <- sum(data[,column])
  }
  return(my_sum)
}

df <- data.frame(x = 1:5,
                 y = 11:15)

# Works    
test2(df, "x")

# Works
df |>
  dplyr::mutate(new_col = test2(df, "x"))

# Doesn't work
df |> 
  dplyr::mutate(across(c("x", "y"), ~test2(df, .)))

Error in `dplyr::mutate()`:
! Problem while computing `..1 = across(c("x", "y"), ~test2(df, .))`.
Caused by error in `across()`:
! Problem while computing column `x`.
Caused by error in `test1()`:
! error in test1
Run `rlang::last_error()` to see where the error occurred.
Warning message:
Problem while computing `..1 = across(c("x", "y"), ~test2(df, .))`.
i the condition has length > 1 and only the first element will be used