Use .data in pack()

[yusuzech]

In dplyr functions, the first argument are mostly .data to prevent possible conflicts. But in tidyr, most functions use data as the first argument. And it creates an issue that when setting data as column name.

For example:

Since data is already used as the first argument of pack() function, when using "data" as a column, it will throw an error message.

df <- tibble(x1 = 1:3, x2 = 4:6, x3 = 7:9, y = 1:3)
df %>% pack(data = c(x1, x2, x3), y = y) 

throws an error message:

Error: All elements of `...` must be named

The same applies to at least pack, unpack, chop and unchop

[dpprdan]

The first argument of nest() has changed from data to .data.
(from tidyr changelog)

🤔 why was this only changed for nest(), but not for unnest(), pack() etc.?

[hadley]

@dpprdan it clearly just slipped my mind

[dpprdan]

And here I was thinking that there is a totally valid reason that I just did not get. 😄

Relatedly in the consistent argument naming department: tibble::tibble(), vctrs::vec_c(), and others have a .name_repair argument, whereas the unnest() function family uses a names_repair argument for the same purpose. (I personally don't mind the (missing) dot so much as the plural 's').

[hadley]

I think this only applies to pack() and unnest(), because chop(), unchop(), and unpack() don't have ....

[hadley]

And ... is deprecated in unnest(), so pack() is the only function that needs a fix.