[Rd] bug in "all.vars()"

[Duncan Murdoch]

On 15/12/2020 7:33 a.m., Daniel Lüdecke wrote:
> Hello,
> I'm not sure if the following is intended, or a bug, but "all.vars()"
> returns different values, depending on whether "TRUE" or "T" is used inside
> "poly()": >
> m1 <- lm(Sepal.Length ~ poly(Sepal.Width, 2, raw = TRUE) + Petal.Length,
> data = iris)
> m2 <- lm(Sepal.Length ~ poly(Sepal.Width, 2, raw = T) + Petal.Length, data =
> iris)
> 
> all.vars(formula(m1))
> #> [1] "Sepal.Length" "Sepal.Width"  "Petal.Length"
> all.vars(formula(m2))
> #> [1] "Sepal.Length" "Sepal.Width"  "T"            "Petal.Length"
> 
> I know that "T" is no reserved word, but rather a global variable.
> Nonetheless, I just wanted to be clear if there is a bug, or everything
> works as intended.


It looks as though it is working as documented.  formula(m2) gives

Sepal.Length ~ poly(Sepal.Width, 2, raw = T) + Petal.Length

and all.vars says it returns "a character vector containing all the 
names which occur in an expression or call."  "T" is a name.

There are good reasons why "R CMD check" warns about using T when you 
mean TRUE.

Duncan Murdoch