allow arrange() to accept NULL argument

[DanChaltiel]

When programming with dplyr, it can be handy to pass NULL as a value.

For instance when using group_by(), I can pass NULL when I want to use my custom function but I don't want to group on any variable.

However, while group_by() allow this syntax, arrange() throws an unexpected error:

dplyr::group_by(iris, NULL)[1,]
#> # A tibble: 1 x 5
#>   Sepal.Length Sepal.Width Petal.Length Petal.Width Species
#>          <dbl>       <dbl>        <dbl>       <dbl> <fct>  
#> 1          5.1         3.5          1.4         0.2 setosa
dplyr::arrange(iris, NULL)[1,]
#> Error in mapply(.f, .x, .y, MoreArgs = list(...), SIMPLIFY = FALSE): zero-length inputs cannot be mixed with those of non-zero length

^{Created on 2022-02-24 by the reprex package (v2.0.1)}

This might be related to #6093.

[hadley]

Looks like this has been fixed:

df <- dplyr::arrange(mtcars, NULL)

^{Created on 2022-07-21 by the reprex package (v2.0.1)}

But there's a new bug if you mix NULL with other variables:

df <- dplyr::arrange(mtcars, NULL, am)
#> Error: `direction` should have length 1 or length equal to the number of columns of `x` when `x` is a data frame.

^{Created on 2022-07-21 by the reprex package (v2.0.1)}

[DavisVaughan]

A correct solution should handle both of these cases:

df <- dplyr::arrange(mtcars, NULL, am)
#> Error: `direction` should have length 1 or length equal to the number of columns of `x` when `x` is a data frame.

x <- NULL
df <- dplyr::arrange(mtcars, x, am)
#> Error: `direction` should have length 1 or length equal to the number of columns of `x` when `x` is a data frame.