regex - R dplyr filter based on matching search term with first words of any work in select columns

Question

I'm trying to filter words from selected columns based on keywords that start the words in the text of match a particular regular expression. Here, I'm trying to pick all words starting with "bio" or "15". But the search terms can also be found in the middle of some words like symbiotic for the Name column and 161540 for the Code column.

**Name**                     **Code**
Biofuel is good          159403
Bioecological is good    161540
Probiotics is good       159883
Good is symbiotic        1877447

I tried the code below

Innov_filter <- Innov_Data %>% 
  select(everything()) %>% 
  filter(str_detect(str_to_lower(Name), "bio") | str_detect(str_to_lower(Code), "bio"))

This is however not working because it is filtering the last row which doesn't fit into any of the conditions. I will appreciate help in strict search based on the first appearance of the search term as part of the word and not just in any location of the word.

Thanks

Answer 1

We could use filter_all with any_vars

df %>% 
   filter_all(any_vars(str_detect(str_to_lower(.), "^(bio|15)")))
#                  Name   Code
#1       Biofuel is good 159403
#2 Bioecological is good 161540
#3    Probiotics is good 159883

NOTE: If it is a subset of columns that needs the conditions to apply, use filter_at

If we need to pick any word that start with 'Bio' in a sentence, wrap with word boundary (\b)

df %>% 
   filter_all(any_vars(str_detect(str_to_lower(.), "\bbio|^15")))

data

df <- structure(list(Name = structure(c(2L, 1L, 4L, 3L), 
   .Label = c("Bioecological is good", 
"Biofuel is good", "Good is symbiotic", "Probiotics is good"), 
  class = "factor"), 
Code = c(159403, 161540, 159883, 1877447)), class = "data.frame", row.names = c(NA, 
 -4L))