如何创建列值为 TRUE 的列名的列表列

Question

我有一些关于多项选择问题的调查数据，这些数据已经输出，每个可能的选择都有一列，填充有 TRUE/FALSE 值，例如

library(dplyr)

dat <- tribble(
  ~name, ~state, ~onions, ~`sweet potatoes`, ~garlic,
  "Tom", "WV", TRUE, FALSE, TRUE,
  "Larry", "NC", FALSE, TRUE, FALSE,
  "Beth", "NY", TRUE, TRUE, TRUE
)

dat

#>   name  state onions `sweet potatoes` garlic
#>   <chr> <chr> <lgl>  <lgl>            <lgl> 
#> 1 Tom   WV    TRUE   FALSE            TRUE  
#> 2 Larry NC    FALSE  TRUE             FALSE 
#> 3 Beth  NY    TRUE   TRUE             TRUE  

如何创建一个列表列，其中包含受访者回答“真”的列名？

类似于https://stackoverflow.com/a/9508203/1009730

我尝试了多种主题的变体

dat %>%
  rowwise() %>%
  mutate(fav_foods = list(names(which(~ is.logical(.x) && .x))))

给出错误：

错误：mutate() 输入 fav_foods 有问题。 'which' 的 x 参数不符合逻辑 i 输入 fav_foods 是 list(names(which(~is.logical(.x) && .x)))。 i 错误发生在第 1 行。

Answer 1

加上purrr，你可以这样做：

dat %>%
 mutate(fav_foods = pmap_chr(across(where(is.logical)), ~ toString(names(c(...))[which(c(...))])))

  name  state onions `sweet potatoes` garlic fav_foods                     
  <chr> <chr> <lgl>  <lgl>            <lgl>  <chr>                         
1 Tom   WV    TRUE   FALSE            TRUE   onions, garlic                
2 Larry NC    FALSE  TRUE             FALSE  sweet potatoes                
3 Beth  NY    TRUE   TRUE             TRUE   onions, sweet potatoes, garlic

或者如果你需要它作为一个列表：

dat %>%
 mutate(fav_foods = pmap(across(where(is.logical)), ~ names(c(...))[which(c(...))]))

  name  state onions `sweet potatoes` garlic fav_foods
  <chr> <chr> <lgl>  <lgl>            <lgl>  <list>   
1 Tom   WV    TRUE   FALSE            TRUE   <chr [2]>
2 Larry NC    FALSE  TRUE             FALSE  <chr [1]>
3 Beth  NY    TRUE   TRUE             TRUE   <chr [3]>

Answer 2

可以使用c_across 更正OP 的代码，而无需使用dplyr 以外的任何其他外部包

library(dplyr)
out <- dat %>% 
       rowwise %>% 
       mutate(fav_foods = list(names(select(.,
            where(is.logical)))[c_across(where(is.logical))])) %>%
       ungroup 
 out
# A tibble: 3 x 6
  name  state onions `sweet potatoes` garlic fav_foods
  <chr> <chr> <lgl>  <lgl>            <lgl>  <list>   
1 Tom   WV    TRUE   FALSE            TRUE   <chr [2]>
2 Larry NC    FALSE  TRUE             FALSE  <chr [1]>
3 Beth  NY    TRUE   TRUE             TRUE   <chr [3]>
> out$fav_foods
[[1]]
[1] "onions" "garlic"

[[2]]
[1] "sweet potatoes"

[[3]]
[1] "onions"         "sweet potatoes" "garlic"        

---

或者我们可以转为“长”格式，filtername 列并将其绑定到原始格式

library(dplyr)
library(tidyr)
dat %>%
  select(where(is.logical)) %>% 
  mutate(rn = row_number()) %>% 
  pivot_longer(cols = -rn) %>%        
  filter(value) %>% group_by(rn) %>%
  summarise(name1 = list(name), .groups = 'drop') %>% 
  select(name1) %>% 
  bind_cols(dat, .)

-输出

# A tibble: 3 x 6
#  name  state onions `sweet potatoes` garlic name1    
#  <chr> <chr> <lgl>  <lgl>            <lgl>  <list>   
#1 Tom   WV    TRUE   FALSE            TRUE   <chr [2]>
#2 Larry NC    FALSE  TRUE             FALSE  <chr [1]>
#3 Beth  NY    TRUE   TRUE             TRUE   <chr [3]>

Answer 3

使用pivot_long 然后summarize

output <- dat %>%
  # get all the columns not name/state into variable & TRUE/FALSE values into values
  pivot_longer(cols = !one_of("name", "state"),
    names_to = "variable", values_to = "value") %>%
  # combined all variables that have values TRUE into variable_true
  # variables that have values FALSE into variable_false
  group_by(name, state) %>%
  arrange(variable) %>%
  summarize(variable_true = paste0(variable[value], collapse = ","),
    variable_false = paste0(variable[!value],collapse = ","),
    .groups = "drop")

这是输出

# A tibble: 3 x 4
  name  state variable_true                variable_false  
  <chr> <chr> <chr>                        <chr>           
1 Beth  NY    garlic,onions,sweet potatoes ""              
2 Larry NC    sweet potatoes               "garlic,onions" 
3 Tom   WV    garlic,onions                "sweet potatoes"

如果您还想拥有以前的列，请添加原始数据

dat %>% left_join(output, by = c("name", "state"))  

你有这个

# A tibble: 3 x 7
  name  state onions `sweet potatoes` garlic variable_true                variable_false  
  <chr> <chr> <lgl>  <lgl>            <lgl>  <chr>                        <chr>           
1 Tom   WV    TRUE   FALSE            TRUE   garlic,onions                "sweet potatoes"
2 Larry NC    FALSE  TRUE             FALSE  sweet potatoes               "garlic,onions" 
3 Beth  NY    TRUE   TRUE             TRUE   garlic,onions,sweet potatoes ""