在R中遍历变量列表并创建一个以变量值为条件的新变量

Question

每当其他12个var（var1至var12）中的任何一个满足许多条件之一时，我都会尝试创建一个新变量（类别）：

for (i in 1:12) {
  new_df <- df %>% mutate(
    category = case_when(df[paste0("var",i)] == "09F" ~ "arts",
                         df[paste0("var",i)] == "09O" ~ "arts_o",
                         TRUE ~ " "))
}

但是category变量仍然为空...有人知道为什么这段代码不起作用吗？我应该使用其他方法吗？

谢谢！

Answer 1

我们可以使用mutate_at将函数应用于多个列，而不是for循环。试试：

library(dplyr)

new_df <- df %>% 
    mutate_at(vars(var1:var12), ~list(category = case_when(. == "09F" ~ "arts",
                                                           . == "09O" ~ "arts_o",
                                                           TRUE ~ " ")))

Answer 2

仅使用一些随机示例数据来获取变量（var1和var2，但可以将其推广到您的情况），根据var1和var2的值为每个对象分配类别的问题可能是像这样解决：

set.seed(42)

# Example data
df <- data.frame(
  id = 1:10,
  var1 = sample(c("09F", "09O", "10F"), 10, replace = TRUE),
  var2 = sample(c("09F", "09O", "10F"), 10, replace = TRUE),
  stringsAsFactors = FALSE
)

library(tidyr)
library(dplyr)

df %>% 
  # Convert to long
  pivot_longer(var1:var2) %>% 
  # Assign category
  group_by(id) %>% 
  mutate(category = case_when(
    any(value == "09F") ~ "arts",
    any(value == "09O") ~ "arts_o",
    TRUE ~ " ")) %>% 
  ungroup() %>% 
  # Convert back to wide
  pivot_wider(id_cols = c(id, category), names_from = "name", values_from = "value")
#> # A tibble: 10 x 4
#>       id category var1  var2 
#>    <int> <chr>    <chr> <chr>
#>  1     1 "arts"   09F   09F  
#>  2     2 "arts"   09F   09F  
#>  3     3 "arts"   09F   09O  
#>  4     4 "arts"   09F   09O  
#>  5     5 "arts_o" 09O   09O  
#>  6     6 "arts_o" 09O   10F  
#>  7     7 "arts_o" 09O   10F  
#>  8     8 "arts"   09F   09F  
#>  9     9 "arts"   10F   09F  
#> 10    10 " "      10F   10F

^{由reprex package（v0.3.0）在2020-04-15创建}

Answer 3

整洁的方法：

set.seed(234)
sample(c("09F", "09O", "xyz"), 12 * 10, replace = TRUE, prob = c(0.05, 0.05, 0.9)) %>%
  matrix(ncol = 12) %>%
  `colnames<-`(str_c("var", 1:12)) %>%
  as_tibble() -> d

d %>%
  mutate(category = map_chr(
    pmap(d %>%
           select(starts_with("var")),
         c),
    ~case_when(any(.x == "09F") ~ "arts",
               any(.x == "09O") ~ "arts_O",
               TRUE ~ " ")))

## A tibble: 10 x 13
#   var1  var2  var3  var4  var5  var6  var7  var8  var9  var10 var11 var12 category
#   <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr> <chr>   
# 1 xyz   xyz   xyz   xyz   xyz   xyz   xyz   09O   xyz   xyz   xyz   xyz   "arts_O"
# 2 xyz   xyz   xyz   xyz   09O   xyz   xyz   xyz   xyz   xyz   xyz   xyz   "arts_O"
# 3 xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   " "     
# 4 xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   " "     
# 5 xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   09O   xyz   xyz   xyz   "arts_O"
# 6 xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   09F   xyz   xyz   "arts"  
# 7 09O   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   "arts_O"
# 8 xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   09F   xyz   xyz   xyz   "arts"  
# 9 09O   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   "arts_O"
#10 xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   xyz   " "