Convert any element that does not start with a specific string to NA-CodePudding

I have a large data frame that looks like df2. I want to convert any element across columns code1, code2 that does not start with AT to NA.

library(tidyverse)

df2 <- tibble(type=c("Jeep", "4x4", "convertible"), 
              code1=c("ATG1",NA, "ATG2"), 
              code2=c("random", "ATG3", "xyz"))
df2
#> # A tibble: 3 × 3
#>   type        code1 code2 
#>   <chr>       <chr> <chr> 
#> 1 Jeep        ATG1  random
#> 2 4x4         <NA>  ATG3  
#> 3 convertible ATG2  xyz

^{Created on 2022-09-29 with reprex v2.0.2}

I want my data to look like this #> type code1 code2 #> #> 1 Jeep ATG1 NA #> 2 4x4 ATG3
#> 3 convertible ATG2 NA

CodePudding user response：

You could do

df2 %>%
  mutate(across(code1:code2, ~ifelse(substr(.x, 1, 2) == 'AT', .x, NA)))
#> # A tibble: 3 x 3
#>   type        code1 code2
#>   <chr>       <chr> <chr>
#> 1 Jeep        ATG1  NA   
#> 2 4x4         NA    ATG3 
#> 3 convertible ATG2  NA

CodePudding user response：

With replace and grepl:

df2 %>% 
  mutate(across(starts_with("code"), ~ replace(.x, !grepl("^AT", .x), NA)))