How to use mutate() across() only for specific rows-CodePudding

The data:

df <- tribble(
  ~name, ~val.I, ~val.V, ~`val.%`,
  "Peter", 123, 12.4, 14,
  "Peter in %", 111, 532, 57,
  "Harald", 2222, 3333, 444,
  "Harald in %", 22, 15, 203,
)

I would like to mutate (function: divide the value by 100) that contain .I and .V, but only for rows that contain the string in % in the column name. I know an easy for-loop solution, but could not figure out it with dplyr. Any hints much appreciated! :)

CodePudding user response：

A dplyr option with mutate and across using matches for the specific columns. You can use the following code:

library(dplyr)

df %>% 
  mutate(across(matches(".I|.V"), ~ if_else(row_number() %in% grep("in %", name), ./100, .)))

Output:

# A tibble: 4 × 4
  name          val.I   val.V `val.%`
  <chr>         <dbl>   <dbl>   <dbl>
1 Peter        123      12.4       14
2 Peter in %     1.11    5.32      57
3 Harald      2222    3333        444
4 Harald in %    0.22    0.15     203

CodePudding user response：

A possible solution in base R:

library(tibble)

df <- tribble(
  ~name, ~val.I, ~val.V, ~`val.%`,
  "Peter", 123, 12.4, 14,
  "Peter in %", 111, 532, 57,
  "Harald", 2222, 3333, 444,
  "Harald in %", 22, 15, 203,
)

df[grepl("in \\%",df$name), grep(".V|.I", names(df))] <- 
       df[grepl("in \\%",df$name), grep(".V|.I", names(df))] /100

#> # A tibble: 4 × 4
#>   name          val.I   val.V `val.%`
#>   <chr>         <dbl>   <dbl>   <dbl>
#> 1 Peter        123      12.4       14
#> 2 Peter in %     1.11    5.32      57
#> 3 Harald      2222    3333        444
#> 4 Harald in %    0.22    0.15     203

With dplyr:

library(dplyr)

df %>% 
 mutate(across(contains(c(".V",".I")), ~ if_else(grepl("in \\%",name), .x/100, .x)))

#> # A tibble: 4 × 4
#>   name          val.I   val.V `val.%`
#>   <chr>         <dbl>   <dbl>   <dbl>
#> 1 Peter        123      12.4       14
#> 2 Peter in %     1.11    5.32      57
#> 3 Harald      2222    3333        444
#> 4 Harald in %    0.22    0.15     203