Creating new column names using dplyr across and .names-CodePudding

I have the following data frame:

df <- data.frame(A_TR1=sample(10:20, 8, replace = TRUE),A_TR2=seq(2, 16, by=2), A_TR3=seq(1, 16, by=2),
                 B_TR1=seq(1, 16, by=2),B_TR2=seq(2, 16, by=2), B_TR3=seq(1, 16, by=2))
> df
  A_TR1 A_TR2 A_TR3 B_TR1 B_TR2 B_TR3
1    11     2     1     1     2     1
2    12     4     3     3     4     3
3    18     6     5     5     6     5
4    11     8     7     7     8     7
5    17    10     9     9    10     9
6    17    12    11    11    12    11
7    14    14    13    13    14    13
8    11    16    15    15    16    15

What I would like to do, is subtract B_TR1 from A_TR1, B_TR2 from A_TR2, and so on and create new columns from these, similar to below:

df$x_TR1 <- (df$A_TR1 - df$B_TR1)
df$x_TR2 <- (df$A_TR2 - df$B_TR2)
df$x_TR3 <- (df$A_TR3 - df$B_TR3)
> df
  A_TR1 A_TR2 A_TR3 B_TR1 B_TR2 B_TR3 x_TR1 x_TR2 x_TR3
1    12     2     1     1     2     1    11     0     0
2    11     4     3     3     4     3     8     0     0
3    19     6     5     5     6     5    14     0     0
4    13     8     7     7     8     7     6     0     0
5    12    10     9     9    10     9     3     0     0
6    16    12    11    11    12    11     5     0     0
7    16    14    13    13    14    13     3     0     0
8    18    16    15    15    16    15     3     0     0

I would like to name these columns "x TR1", "x TR2", etc. I tried to do the following:

xdf <- df%>%mutate(across(starts_with("A_TR"), -across(starts_with("B_TR")), .names="x TR{.col}"))

However, I get an error in mutate():

attempt to select less than one element in integerOneIndex

I also don't know how to create the proper column names, in terms of getting the numbers right -- I am not even sure the glue() syntax allows for it. Any help appreciated here.

CodePudding user response：

We could use .names in the first across to replace the substring 'a' with 'x' from the column names (.col) while subtracting from the second set of columns

library(dplyr)
library(stringr)
df <- df  %>% 
   mutate(across(starts_with("A_TR"), 
     .names = "{str_replace(.col, 'A', 'x')}") - 
       across(starts_with("B_TR")))

-output

df
   A_TR1 A_TR2 A_TR3 B_TR1 B_TR2 B_TR3 x_TR1 x_TR2 x_TR3
1    10     2     1     1     2     1     9     0     0
2    10     4     3     3     4     3     7     0     0
3    16     6     5     5     6     5    11     0     0
4    12     8     7     7     8     7     5     0     0
5    20    10     9     9    10     9    11     0     0
6    19    12    11    11    12    11     8     0     0
7    17    14    13    13    14    13     4     0     0
8    14    16    15    15    16    15    -1     0     0