[R] odd feature

[Gabor Grothendieck]

On 5/22/06, Martin Maechler <maechler at stat.math.ethz.ch> wrote:
> >>>>> "TL" == Thomas Lumley <tlumley at u.washington.edu>
> >>>>>     on Mon, 22 May 2006 07:09:09 -0700 (PDT) writes:
>
>    TL> On Mon, 22 May 2006, Gabor Grothendieck wrote:
>    >> Due to lazy evaluation, I don't think a and b are fully evaluated:
>    >>
>    >>> ifelse(1, a <- 1, b <- 2)
>    >> [1] 1
>    >>> a
>    >> [1] 1
>    >>> b
>    >> Error: object "b" not found
>
>    TL> yes. If you look at the code for ifelse() it evaluates
>    TL> the second argument if any test values are TRUE and the
>    TL> third argument if any test values are FALSE, so in the
>    TL> scalar case it does not evaluate both arguments.
>
> yes.
> And (Gabor), if you recall what I said earlier,  it's pretty clear
> that I would not even *consider* using ifelse() in the scalar
> case.  So my original statement was about the case ifelse() is
> designed for: The non-trivial vector case, and there my
> statement of "fully evaluated"  does apply.
>
> Martin
>

I don't think they are fully evaluated in the vector case either:

> ifelse(rep(TRUE,4), a <- 1:4, b <- 1:4)
[1] 1 2 3 4
> a
[1] 1 2 3 4
> b
Error: object "b" not found