Improve error message when pivot_longer() is applied on data with a "name" column

[DanChaltiel]

Hi,

When you apply pivot_longer() on a dataframe that already has a column named name, the error message is confusing:

library(tidyverse)
tibble(name=c("x", "y"), a=c("a", "b"), b=c("c", "d")) %>% 
  pivot_longer(-name)
#> Error in `pivot_longer()`:
#> ! Names must be unique.
#> x These names are duplicated:
#>   * "name" at locations 1 and 2.
#> i Use argument `names_repair` to specify repair strategy.

^{Created on 2023-04-14 with reprex v2.0.2}

Wouldn't it be possible to check if data has a column named like names_to and hint the user to use names_to to correct the problem instead of names_repair which doesn't help much here?

[hadley]

A bit like #1367, but I think the root cause is the default values for names_to. I'm not sure what we should do about this.

[DanChaltiel]

Here is how I would somehow formulate it:

library(tidyverse)

pivot_longer2 = function(data, cols, names_to = "name", ...){
  wrong_names = names(data)[names(data) %in% names_to]
  if(length(wrong_names)>0){
    cli::cli_abort("{.arg names_to}={.val {wrong_names}} is already used in {.arg data} column names.")
  }
  pivot_longer(data=data, cols={{cols}}, names_to=names_to, ...)
}

tibble(name=c("x", "y"), a=c("a", "b"), b=c("c", "d")) %>% 
  pivot_longer2(-name)
#> Error in `pivot_longer2()`:
#> ! `names_to`="name" is already used in `data` column names.
tibble(name2=c("x", "y"), a=c("a", "b"), b=c("c", "d")) %>% 
  pivot_longer2(-name2)
#> # A tibble: 4 x 3
#>   name2 name  value
#>   <chr> <chr> <chr>
#> 1 x     a     a    
#> 2 x     b     c    
#> 3 y     a     b    
#> 4 y     b     d

^{Created on 2023-11-02 with reprex v2.0.2}

[jpiaskowski]

One solution could be including a note in the documentation that if you have a column called "name" it will result in an error. At least that tells users what the problem is so we spend less time scratching our heads, wondering what we did wrong. (this just happened to me today, fortunately renaming the column is a very easy fix)