dplyr - ifelse 在 mutate 内使用列表

给定一个数据框df像下面这样

text <- "
model,var,value
M1,a,12211
M1,b1,10.21
M1,b2,5.07
M1,c1,41.8
M1,c2,58.2
M1,d,1.6
M2,a,11922
M2,b1,15.6
M2,b2,8.9
M2,c1,38.1
M2,c2,61.9
M2,d,1.8
M2,a,13101
M2,b1,9.21
M2,b2,6.56
M2,c1,36.07
M2,c2,63.93
M2,d,1.75
"
df <- read.table(textConnection(text), sep=",", header = T)

我想添加一个列var2基于的价值var通过 dplyrmutate.

基于以下逻辑。

If var == 'b1' or var == 'b2' then All B If var == 'c1' or var == 'c2' then All C else var

我想按如下方式存储映射并使用它来构建上述逻辑

mapping <- c("All B"= list(c('b1', 'b2')), "All C" = list(c('c1', 'c2')))
> mapping    
$`All B`
[1] "b1" "b2"

$`All C`
[1] "c1" "c2"

预期输出是

   model var    value var2
1     M1   a 12211.00    a
2     M1  b1    10.21    All B
3     M1  b2     5.07    All B
4     M1  c1    41.80    All C
5     M1  c2    58.20    All C
6     M1   d     1.60    d
7     M2   a 11922.00    a
8     M2  b1    15.60    All B
9     M2  b2     8.90    All B
10    M2  c1    38.10    All C
11    M2  c2    61.90    All C
12    M2   d     1.80    d
13    M2   a 13101.00    a
14    M2  b1     9.21    All B
15    M2  b2     6.56    All B
16    M2  c1    36.07    All C
17    M2  c2    63.93    All C
18    M2   d     1.75    d

我计划将 dplyr 与ifelse如下

df %>%
  mutate(var2 = ifelse(# what should go here )

我们可以创建一个返回列表名称的函数，如果var存在于mapping否则它返回var。我们可以用rowwise()为每一行执行此函数。

get_right_mapping <- function(var) {
   names(which(sapply(mapping, function(x) var %in% x)))
 }

library(dplyr)
df %>%
   rowwise() %>%
   mutate(var2 = ifelse(var %in% unlist(mapping), get_right_mapping(var), var))


#   model var      value  var2 
#   <fct> <chr>    <dbl> <chr>
# 1 M1    a        12211    a    
# 2 M1    b1       10.2  All B
# 3 M1    b2        5.07 All B
# 4 M1    c1       41.8  All C
# 5 M1    c2       58.2  All C
# 6 M1    d        1.60     d    
# 7 M2    a        11922    a    
# 8 M2    b1       15.6  All B
# 9 M2    b2        8.90 All B
#10 M2    c1       38.1  All C
#11 M2    c2       61.9  All C
#12 M2    d         1.80 d    
#13 M2    a         13101    a    
#14 M2    b1        9.21 All B
#15 M2    b2        6.56 All B
#16 M2    c1       36.1  All C
#17 M2    c2       63.9  All C
#18 M2    d         1.75 d    

data

mapping <- c( "All A"= list(c('a1', 'a2')), "All B" = list(c('b1', 'b2')), 
              "All C" = list(c('c1','c2')))
df$var <- as.character(df$var)