[Rd] Bounds error when indexing a matrix with NA_character_

[Milan Bouchet-Valat]

Hi!


I've just got a bounds error when trying to index a matrix with
NA_character_:

> y <- matrix(1:4, 2, 2, dimnames = list(A=c("a", "b"), B=c("a", "b")))
> y[NA,]
      B
A       a  b
  <NA> NA NA
  <NA> NA NA
> y[NA_integer_,]
 a 
 b 
NA NA 
> y[NA_character_,]
Error in y[NA_character_, ] : index out of
bounds


Yet, I would have sworn it behaves like integer NA (which apparently
isn't true since it also occurs with R 3.0). Assuming this isn't simply
a bug, what's the rationale behind this, and is it documented anywhere?
For example, ?Extract says:
>      When extracting, a numerical, logical or character ‘NA’ index
>      picks an unknown element and so returns ‘NA’ in the corresponding
>      element of a logical, integer, numeric, complex or character
>      result, and ‘NULL’ for a list.  (It returns ‘00’ for a raw
>      result.)

An indeed that's how it works for vectors:
> x <- c(a=1)
> x[NA_character_]
<NA> 
  NA 

The inconsistency between NA classes, and between vectors and matrices
is surprising.


Thanks for your comments.