使用因素：在 R 中重新排序、重新标记和重新编码

Question

我遇到了一些麻烦 "factoring" 字符变量和 labeling/recoding 不同的级别。使用 tidyverse 或 Base R 例如 recode 或 fct_collapse 等是否有一种有效的方法（即更少的编码）......谢谢

#This is what I have (a character variable)
x <- c("No", "Yes", "No2", "No3", "Maybe", "undecided", 
       "probably", "dont know", NA)
x
#I want a factor with three ordered levels as follows:

#where No = c("No", "No2", "No3")
#Yes = c("Yes")
#other = c("Maybe", "undecided", "probably")
#NA = c("dont know", NA)
# and the levels would be 0 = "No",  1 = "Yes" and 2 = "Maybe"
#that is:
#xfact
# [1] No    Yes   other
# Levels: No Yes other
#
# as.integer(xfact)
# [1] 0, 1, 2```

Answer 1

应该这样做：

library(tidyverse)

x <- c("No", "Yes", "No2", "No3", "Maybe", "undecided", 
       "probably", "dont know", NA)

na_if(x, "dont know") %>% 
  fct_collapse(
    no = c("No", "No2", "No3"),
    yes = c("Yes"),
    other = c("Maybe", "undecided", "probably")
  ) %>% 
  fct_inorder()
#> [1] no    yes   no    no    other other other <NA>  <NA> 
#> Levels: no yes other

^{由 reprex package (v0.3.0)}

于 2020 年 1 月 13 日创建

使用因素：在 R 中重新排序、重新标记和重新编码

working with factors: reordering, relabeling and recoding in R

r

recode

tidyverse

forcats