根据需要重新标记因子水平答案

【问题标题】：Relabel levels of factors as desired根据需要重新标记因子水平
【发布时间】：2022-11-02 20:09:47
【问题描述】：

我想重新标记因素水平如下：

i. If the level is of length 3 and above relabel it to sentence case otherwise do title case

即home doing nothing变成Home doing nothing，yes变成Yes和good practice变成Good Practice

有没有办法做到这一点？

library(tidyverse)

vars <- c("a", "b", "c", "d", "e")


mydata <- tribble(
  ~"a", ~"b", ~"c", ~"d", ~"e", ~"id",
  "yes", "school in Kenya", "r", 10, "good practice", 1,
  "no", "home doing nothing", "python", 12, "doing well", 3,
  "no", "school in Tanzania", "c++", 35, "by walking", 4,
  "yes", "home practising", "c", 65, "practising everyday", 5,
  "no", "home", "java", 78, "sitting alone", 7
) %>%
  mutate(across(.cols = vars, ~as_factor(.)))


# mydata %>%
#   mutate(across(where(is.factor), ~fct_relabel(., str_to_sentence(.))))

【问题讨论】：

标签： r dplyr forcats

【解决方案1】：

这是一种可能的解决方案。请注意，您考虑为factor 的列实际上是character 变量，而不是factor。

library(dplyr)
library(stringr)

mydata %>%
  mutate(across(where(is.character), 
                ~ if_else(stringi::stri_count_words(.x)>=3, 
                          sub("^([a-z])", "\U\1\E", .x, perl=T), 
                          str_to_title(.x))))

# A tibble: 5 x 6
  a     b                  c          d e                      id
  <chr> <chr>              <chr>  <dbl> <chr>               <dbl>
1 Yes   School in Kenya    R         10 Good Practice           1
2 No    Home doing nothing Python    12 Doing Well              3
3 No    School in Tanzania C++       35 By Walking              4
4 Yes   Home Practising    C         65 Practising Everyday     5
5 No    Home               Java      78 Sitting Alone           7

【讨论】：

【解决方案2】：

这里有两个选项。 1）您分别评估 >3 和 <3。或者 2) 你同时评估两者。两者在可读性方面各有利弊。

library(tidyverse)

#option 1
mydata |>
  mutate(across(where((x) length(levels(x)) >= 3), 
              (x) fct_relabel(x, str_to_sentence)),
         across(where((x) length(levels(x)) < 3), 
              (x) fct_relabel(x, str_to_title)))
#> # A tibble: 5 x 6
#>   a     b                  c      d     e                   id   
#>   <fct> <fct>              <fct>  <fct> <fct>               <fct>
#> 1 Yes   School in kenya    R      10    Good practice       1    
#> 2 No    Home doing nothing Python 12    Doing well          3    
#> 3 No    School in tanzania C++    35    By walking          4    
#> 4 Yes   Home practising    C      65    Practising everyday 5    
#> 5 No    Home               Java   78    Sitting alone       7

#option 2
mydata |>
  mutate(across(everything(), (x) if_else(
    rep(length(levels(x)) >= 3, length(x)), 
    fct_relabel(x, str_to_sentence),
    fct_relabel(x, str_to_title)
  )))
#> # A tibble: 5 x 6
#>   a     b                  c      d     e                   id   
#>   <fct> <fct>              <fct>  <fct> <fct>               <fct>
#> 1 Yes   School in kenya    R      10    Good practice       1    
#> 2 No    Home doing nothing Python 12    Doing well          3    
#> 3 No    School in tanzania C++    35    By walking          4    
#> 4 Yes   Home practising    C      65    Practising everyday 5    
#> 5 No    Home               Java   78    Sitting alone       7

【讨论】：