根据 Tidyverse 中的字符串模式创建具有多个类别的列

Question

I checked this question，但不确定如何将其用于多个类别（不仅仅是两个）。 This is conceptually similar，但不确定它是否是字符串的好选择。

我有一个数据框

Gender          Frequency 
female            49719         
male              14835         
NA                712           
female, male      518   

此外，还有更多选项，例如 female, female, female 或 male, male, female。我有几十种组合。

我想创建一个新专栏，其中只有四个类别 - male、female、both、NA。例如，如果一个挡板female 重复多次，则将其归类为female。如果它是不同性别的组合（任何长度） - 称之为both。

期望的输出：

Gender          Frequency      Category
female            49719          female 
male              14835          male
NA                712            NA
female, male      518            both
male, male, male  100            male
male, male, female 100           both

我将不胜感激tidyverse 解决方案。

Answer 1

library(tidyverse)

df %>%
  mutate(row = row_number()) %>%
  separate_rows(Gender, sep = ',\\s*') %>%
  group_by(row) %>%
  summarise(Category = if(n_distinct(Gender) > 1) 'both' else unique(Gender), 
            Gender = toString(Gender), 
            Frequency = first(Frequency)) %>%
  select(Gender, Frequency, Category)

#  Gender             Frequency Category
#  <chr>                  <int> <chr>   
#1 female                 49719 female  
#2 male                   14835 male    
#3 NA                       712 NA      
#4 female, male             518 both    
#5 male, male, male         100 male    
#6 male, male, female       100 both    

或者

df %>%
  mutate(Category = map_chr(str_split(Gender, ',\\s*'), 
                          ~if(n_distinct(.) > 1) 'both' else unique(.)))

可以在基础R中翻译-

df$Category <- sapply(strsplit(df$Gender, ',\\s*'), function(x) 
                   if(length(unique(x)) > 1) 'both' else unique(x))