CODEWITHSUNDEEP
rdplyrstringr
zesla
zesla

Reputation: 11793

how to make str_detect case insensitive within case_when in r

I have a dataframe like below:

dfm = data.frame (names = c('email', 'Facebook', 'walmart', 'target', 'instagram', 'costco'))

I need to create a new column source. When names are facebook, instagram or email, source is media. When names are costco,walmart or target, the source is store.

I use case_when and str_detect. I need to the str_detect to be case insensitive. So below is my code. 

dfm %>% 
  mutate(source = case_when( str_detect(names, fixed('email|facebook|instagram', ignore_case = T))~'media',
                             str_detect(names, 'walmart|costco|target')~ 'store'))

I got:

names       source
email       NA          
Facebook    NA          
walmart     store           
target      store           
instagram   NA          
costco      store

I do not understand why it did not work. Does anyone know why?

I tried the code below, it returns TRUE

str_detect('Facebook', fixed('facebook', ignore_case = T))

Upvotes: 1

Views: 2759

Answers (2)

sometimes_sci
sometimes_sci

Reputation: 199

Also try regex instead of fixed:

dfm %>% 
  mutate(source = case_when( str_detect(names, regex('email|facebook|instagram', ignore_case = T))~'media',
                             str_detect(names, 'walmart|costco|target')~ 'store'))

Upvotes: 4

Ronak Shah
Ronak Shah

Reputation: 388862

When you are using fixed it will not recognize | as regex. If you want to do exact match use word boundaries i.e \\b. 

library(dplyr)
library(stringr)

dfm %>% 
   mutate(source = case_when(str_detect(names, 
                 regex('\\bemail\\b|\\bfacebook\\b|\\binstagram\\b', 
                        ignore_case = TRUE))~'media',
                  str_detect(names, 'walmart|costco|target')~ 'store'))

#      names source
#1     email  media
#2  Facebook  media
#3   walmart  store
#4    target  store
#5 instagram  media
#6    costco  store

Upvotes: 4

Related Questions