替换引用命名向量的整个字符串

Replace whole string referencing a named vector

我想对命名向量进行部分字符串匹配,在匹配的地方使用命名向量作为整个字符串值。

感谢您的帮助。

library(tidyr)
library(stringr)
map <- tibble(str = c('rolex','mazda','nike'),
              cat = c('watch','car','shoe'))
df <- tibble(data = c('I love driving mazda','He wears a rolex to work','awesome nike shoe'))

样本table

> map
# A tibble: 3 x 2
  str   cat  
  <chr> <chr>
1 rolex watch
2 mazda car  
3 nike  shoe 
> df
# A tibble: 3 x 1
  data                    
  <chr>                   
1 I love driving mazda    
2 He wears a rolex to work
3 awesome nike shoe

想要结果

> df
# A tibble: 3 x 2
  data                     cat  
  <chr>                    <chr>
1 I love driving mazda     car  
2 He wears a rolex to work watch
3 awesome nike shoe        shoe 

试过了,但它替换了它找到的字符串,但没有替换上面引用命名向量的整个字符串。

map_v = as.character(map$cat)
names(map_v) = map$str

不需要的结果

> df %>% mutate(cat = str_replace_all(data,map_v))
# A tibble: 3 x 2
  data                     cat                     
  <chr>                    <chr>                   
1 I love driving mazda     I love driving car      
2 He wears a rolex to work He wears a watch to work
3 awesome nike shoe        awesome shoe shoe
library(purrr)
library(dplyr)
library(stringr)

df %>% 
  mutate(cat = map_chr(data, ~ map$cat[str_detect(.x, map$str)]))
# # A tibble: 3 x 2
# data                     cat  
# <chr>                    <chr>
# 1 I love driving mazda     car  
# 2 He wears a rolex to work watch
# 3 awesome nike shoe        shoe 

或者你可以使用 fuzzyjoin 包

library(fuzzyjoin)

df %>% 
  regex_inner_join(map, by = c(data = 'str'))

# # A tibble: 3 x 3
# data                     str   cat  
# <chr>                    <chr> <chr>
# 1 I love driving mazda     mazda car  
# 2 He wears a rolex to work rolex watch
# 3 awesome nike shoe        nike  shoe 

使用 outer + grepl + max.col

的基础 R 选项
df$cat <- with(map, cat[max.col(t(outer(map$str, df$data, Vectorize(grepl))))])

给予

> df
# A tibble: 3 x 2
  data                     cat
  <chr>                    <chr>
1 I love driving mazda     car
2 He wears a rolex to work watch
3 awesome nike shoe        shoe