On 13-01-09 5:03 AM, Berry Boessenkool wrote:>
>
> Hi,
>
> I'm writing a function that needs the input names (as characterstrings)
as part of the output.
> With deparse(substitute( ) ) that works fine, until I replace all zeros
with 0.001 (log is calculated at some time):
>
> tf <- function(input) { input[input==0] <- 0.001 ;
deparse(substitute(input)) }
> myguess <- 42
> tf(myguess) # not "myguess", but "42"
>
>
> Now when I extract the input names before replacing the zeros, this works:
>
> tf <- function(input) { out <- deparse(substitute(input)) ;
input[input==0] <- 0.001 ; out }
> tf(myguess) # correct: "myguess"
> myguess <- 0 ; tf(myguess) # ditto
>
>
> While I did find a workaround, I'm still wondering why this happens.
> Any hints on where to start reading?
Probably the R Language definition, section 2.1.8. The basic
explanation for the behaviour you see is that deparse(substitute(input))
acts on the "input" promise object, looking at its expression slot.
It's
not doing any magic examination of the context in which it was
originally defined.
Once you modify it, it is no longer a promise, and so it has no
expression slot.
Duncan Murdoch