row.name 使用 `structure` 函数作为 c(NA, *integer*)

Question

Does anyone know why when I run this:

row.names(structure(list(speed = c(4, 7), dist = c(2, 22)), 
                    row.names = c(NA, 2L), class = "data.frame"))

I get this:

# "1" "2"

and not c(NA, 2L) ? I mean what row.names argument in structure exactly does to the argument?

I came across this when I tried to use dput to see the structure of some dataframes. eg

dput(cars)

And I noticed the row.names argument in it, which is: c(NA, -50L) .

Answer 1

c(NA, n) is how data frames internally store the row names in the common case of 1:n so as to save space and processing time. This is not supposed to be accessible to the user who is to regard it as "1", "2", ... so the accessor functions translate it.