`select` renames columns if vector of column names is named

[joshua-theisen]

When I use a named vector to select columns, where the column name is specified by a named element in the vector, those columns are renamed. For example:

> df1 <- tibble::tibble(a = 1:4, b = letters[11:14])
> df1
# A tibble: 4 × 2
      a b    
  <int> <chr>
1     1 k    
2     2 l    
3     3 m    
4     4 n    
> col_sel <- c("a", b_but_renamed = "b")
> col_sel
              b_but_renamed 
          "a"           "b" 
> df2 <- df1 %>% dplyr::select(tidyselect::all_of(col_sel))
> df2
# A tibble: 4 × 2
      a b_but_renamed
  <int> <chr>        
1     1 k            
2     2 l            
3     3 m            
4     4 n            

Is this behavior expected? If so, I can't find it in the documentation (though this may be due to my failure to find it). If this behavior is expected, is there an argument to dplyr::select that would allow me to disable this feature, or do I need to always call unname on col_sel to ensure this doesn't happen? Thanks!

[DavisVaughan]

It is expected, and is not disable-able, and I don't think we would add one to do so.

It doesn't look like it is documented right in select()'s help page, so we should probably add that.

I'm surprised we also don't document it in tidyselect, so there are definitely some docs missing:
https://tidyselect.r-lib.org/reference/all_of.html

[DavisVaughan]

We should add documentation to select() about how to use this feature. Here is a simple example:

library(dplyr)

df <- tibble(
  x = 1:4,
  y = 5:8
)

selection <- c(
  new_x = "x",
  new_y = "y"
)

df |> select(all_of(selection))
#> # A tibble: 4 × 2
#>   new_x new_y
#>   <int> <int>
#> 1     1     5
#> 2     2     6
#> 3     3     7
#> 4     4     8