用于循环列表的自定义函数

我有一个有效的自定义函数，但不确定如何让它与输入列表循环。看起来我需要了解 apply() 等，但我对当前的设置还不太了解。该函数使用 rollapply() 查找给定时间范围内的最大指标。

library(zoo)
library(dplyr)

# Data
set.seed(1)
df <- tibble(player = rep(LETTERS[1:2], each = 10),
             minute = rep(1:10, times = 2),
             tdc = sample(100:200,size = 20),
             sumad = sample(1:10, size = 20, replace = TRUE))

# Custom function
x_min_roll <- function(df, metric, n_minutes, fun){
  metric <- ensym(metric)
  newname <- glue::glue("{rlang::as_string(metric)}_x{as.character(n_minutes)}")
  df %>% 
    # dynamically create new column name based on input
    mutate("{newname}" := rollapply(!!metric, n_minutes, fun, align='left', fill=NA)) %>% 
    group_by(player) %>% 
    slice_max(.data[[newname]]) %>% 
    select(player, .data[[newname]])
}

# This works
df %>% 
  x_min_roll(metric = tdc, n_minutes = 2, fun = sum)

# A tibble: 2 x 2
# Groups:   player [2]
  player tdc_x2
  <chr>   <int>
1 A         339
2 B         380

我希望能够做到这一点：

metric_list <- c('tdc', 'sumad')
minutes_list <- c(2,5)

df %>% 
  x_min_roll(metric = metric_list, n_minutes = minutes_list, fun = sum) %>% 
  # maybe a few more steps here.... to get this

# A tibble: 2 x 5
  player tdc_x2 tdc_x5 sumad_x2 sumad_x5
  <chr>   <dbl>  <dbl>    <dbl>    <dbl>
1 A         339    793       20       36
2 B         380    866       19       41

我们可以用map2循环两个向量的相应元素

library(purrr)
library(dplyr)
map2(metric_list, minutes_list, 
  ~ df %>%
    x_min_roll(metric = !!.x, n_minutes = .y, fun = sum))

-output

[[1]]
# A tibble: 2 × 2
# Groups:   player [2]
  player tdc_x2
  <chr>   <int>
1 A         339
2 B         380

[[2]]
# A tibble: 3 × 2
# Groups:   player [2]
  player sumad_x5
  <chr>     <int>
1 A            36
2 B            41
3 B            41

编辑：基于@Onyambu的评论

如果我们想要每个组合，那么使用crossing创建组合

library(tidyr)
crossing(metric_list, minutes_list) %>% 
 pmap(~ df %>% 
      x_min_roll(metric = !!.x, n_minutes = .y, fun = sum))

根据OP的评论，如果我们想合并数据集

crossing(metric_list, minutes_list) %>% 
 pmap(~ df %>% x_min_roll(metric = !!.x, n_minutes = .y, fun = sum)) %>%
    reduce(inner_join, by = 'player')