替换匹配多个不同模式的列值

Question

在下面的数据框中:

library(tidyverse)
df <- tibble(notes=c("Positive result","Negative","NEG","POS >2","pOS","Cannot Determine","2.4","3.1","0.2"))

  notes           
  <chr>           
1 Positive result 
2 Negative        
3 NEG             
4 POS >2          
5 pOS             
6 Cannot Determine
7 2.4             
8 3.1             
9 0.2

我想定义一个单行代码来替换注释列中与模式匹配的条目。如果只有两个条件，我会使用三元运算符。但这里我有 5 个。

我希望将注释中的值替换为:

1. 可以变成双数 -> "3"
2. grepl("pos",tolower(notes)) -> "2"
3. grepl("neg",tolower(notes)) -> "1"
4. 以上都不是 -> “0”

我最初做了:

df %>%
  mutate(notes=ifelse(grepl("[[:digit:]]+",notes)),"3",notes) %>%  # could be coerced into a double
  mutate(notes=ifelse(grepl("pos",tolower(notes))),"2",notes) %>%  # contains "pos"
  mutate(notes=ifelse(grepl("neg",tolower(notes))),"1",notes) %>%  # contains "neg"
  mutate(notes=ifelse(grepl("3|2|1",tolower(notes))),notes,"0") %>%  # none of the above
  type.convert()

期望的输出

notes           
<dbl>           
1 2 
2 1        
3 1             
4 2          
5 2             
6 0
7 3            
8 3             
9 3             

Answer 1

我们可以使用case_when

library(dplyr)
library(stringr)
df %>% 
  mutate(notes1 = toupper(substr(notes, 1, 3)), 
        notes =case_when(notes1 == "POS" ~ 2, 
                         notes1 == 'NEG' ~ 1, 
                         str_detect(notes, '^[0-9.]+$')~ 3,
                         TRUE ~ 0)) %>%
  select(-notes1)
# A tibble: 9 x 1
#  notes
#  <dbl>
#1     2
#2     1
#3     1
#4     2
#5     2
#6     0
#7     3
#8     3
#9     3

---

如果我们需要保持数值不变，一种选择是 as.numeric 然后是 coalesce

df %>% 
 mutate(notes1 = toupper(substr(notes, 1, 3)), 
      notes2 =case_when(notes1 == "POS" ~ 2, 
                       notes1 == 'NEG' ~ 1, 
                       str_detect(notes, '^[0-9.]+$')~ 3,
                       TRUE ~ 0)) %>%
 select(-notes1) %>% 
 mutate(notes = coalesce(as.numeric(notes), notes2)) %>% 
 select(-notes2)