r - 如何使用 tidyverse 函数在 R 中的嵌套 df 中选择所有共享列？

Question

我在下面提出了一个代表。我创建了一个嵌套的df. 该cols列表示嵌套中的列df。我正在使用该cols列传递给使用map.

在这种情况下，这将是lifeExp列。

我如何传递一个select查看cols列和selects该行中存在的字符向量列名称的a，然后是嵌套dfselect中的该行？sd_col另外，我想传递country,continent和sd同一个select. 所以在这种情况下，我们将选择 , country, continent, lifeExp, 和sd嵌套df在sd_col.

我猜这可以使用purrr.

library(gapminder)
library(tidyverse)

data <- gapminder_unfiltered


sample_fn1 <- function(data, cols){
  testdata <- data %>% 
    mutate(across(!!cols, mean))
}


sample_fn2 <- function(data, cols){
  testdata <- data %>% 
    mutate(sd = sd(pop))
}

nested_data <- data %>% 
  filter(country %in% c("United States", "Mexico", "Canada", "Argentina", "Brazil", "Italy", "Japan")) %>% 
  mutate(
    country_group = case_when(
    country %in% c("United States", "Mexico", "Canada") ~ "North America",
    country %in% c("Argentina", "Brazil") ~ "South America",
    country %in% c("Italy", "Japan") ~ "Eurasia"
  )
  ) %>% 
  group_by(country_group) %>% 
  nest() %>% 
  ungroup()


> nested_data 
# A tibble: 3 x 2
  country_group data                  
  <chr>         <list>                
1 South America <tibble[,6] [24 x 6]> 
2 North America <tibble[,6] [127 x 6]>
3 Eurasia       <tibble[,6] [114 x 6]>

下一个

使用的列

col_tbl <- tibble::tribble(
  ~country_group, ~cols,
  "North America", c("lifeExp", "pop"),
  "South America", c("lifeExp", "gdpPercap"),
  "Eurasia"       , c("lifeExp", "pop")
)

> col_tbl
# A tibble: 3 x 2
  country_group cols     
  <chr>         <list>   
1 North America <chr [2]>
2 South America <chr [2]>
3 Eurasia       <chr [2]>

加入 cols 以用于 df

nested_data <- nested_data %>% 
  left_join(col_tbl)
#> Joining, by = "country_group"

> nested_data
# A tibble: 3 x 3
  country_group data                   cols     
  <chr>         <list>                 <list>   
1 South America <tibble[,6] [24 x 6]>  <chr [2]>
2 North America <tibble[,6] [127 x 6]> <chr [2]>
3 Eurasia       <tibble[,6] [114 x 6]> <chr [2]>

运行 2 个函数

nested_data <- nested_data %>% 
  mutate(nest1 = map2(data, cols, sample_fn1),
         nest2 = map(data, sample_fn2),
         # select specified columns 
         nest1 = map(nest1, ~select(.x, c(country, year, continent)))
         )

> nested_data
# A tibble: 3 x 5
  country_group data                   cols      nest1                  nest2                 
  <chr>         <list>                 <list>    <list>                 <list>                
1 South America <tibble[,6] [24 x 6]>  <chr [2]> <tibble[,3] [24 x 3]>  <tibble[,7] [24 x 7]> 
2 North America <tibble[,6] [127 x 6]> <chr [2]> <tibble[,3] [127 x 3]> <tibble[,7] [127 x 7]>
3 Eurasia       <tibble[,6] [114 x 6]> <chr [2]> <tibble[,3] [114 x 3]> <tibble[,7] [114 x 7]>

标签： rtidyrpurrr