r - 使用 tidyeval 在自制函数中选择函数

Question

所以这个例子基本上来自https://tidyeval.tidyverse.org/dplyr.html#patterns-for-single-arguments它工作得很好:

library(tidyverse)
group_mean <- function(df, group_var, summary_var){
  group_var <-  rlang::enquo(group_var)
  summary_var <-rlang::enquo(summary_var)

  name <- paste0(rlang::quo_name(summary_var), "_mean")

  df %>%
    dplyr::group_by(!!group_var) %>%
    dplyr::summarise(!!name := mean(!!summary_var, na.rm = TRUE))
}

mtcars %>% group_mean(group_var = cyl, summary_var = disp)
#> # A tibble: 3 x 2
#>     cyl disp_mean
#>   <dbl>     <dbl>
#> 1     4      105.
#> 2     6      183.
#> 3     8      353.

我想，例如有时能够选择中位数而不是平均值，例如将函数名称更改为 group_stat()。

Answer 1

你可以这样做。我不太清楚这是如何工作的，但我已经在 library(purrr) 的源代码中看到这种方法用于 as_mapper():

https://github.com/tidyverse/purrr/blob/master/R/as_mapper.R

library(tidyverse)

group_stat <- function(df, group_var, summary_var, .f) {

    func <- rlang::as_closure(.f)

    group_var <-  rlang::enquo(group_var)
    summary_var <-rlang::enquo(summary_var)

    name <- paste0(rlang::quo_name(summary_var), "_", deparse(substitute(.f)))

    df %>%
        dplyr::group_by(!!group_var) %>%
        dplyr::summarise(!!name := func(!!summary_var, na.rm = TRUE))
}

mtcars %>% 
    group_stat(group_var = cyl, summary_var = disp, median)
#> # A tibble: 3 x 2
#>     cyl disp_median
#>   <dbl>       <dbl>
#> 1     4        108 
#> 2     6        168.
#> 3     8        350.

mtcars %>% 
    group_stat(group_var = cyl, summary_var = disp, mean)
#> # A tibble: 3 x 2
#>     cyl disp_mean
#>   <dbl>     <dbl>
#> 1     4      105.
#> 2     6      183.
#> 3     8      353.

mtcars %>% 
    group_stat(group_var = cyl, summary_var = disp, max)
#> # A tibble: 3 x 2
#>     cyl disp_max
#>   <dbl>    <dbl>
#> 1     4     147.
#> 2     6     258 
#> 3     8     472

mtcars %>% 
    group_stat(group_var = cyl, summary_var = disp, min)
#> # A tibble: 3 x 2
#>     cyl disp_min
#>   <dbl>    <dbl>
#> 1     4     71.1
#> 2     6    145  
#> 3     8    276.

^{由 reprex package 创建于 2019-05-02 (v0.2.1)}