[Rd] mapply(): Special case of USE.NAMES=TRUE with recent R-devel updates

2021-11-30 Thread Henrik Bengtsson
Hi,

in R-devel (4.2.0), we now get:

> mapply(paste, "A", character(), USE.NAMES = TRUE)
named list()

Now, in ?mapply we have:

USE.NAMES: logical; use the names of the first ... argument, or if
that is an unnamed character vector, use that vector as the names.

This basically says we should get:

> answer <- list()
> first <- "A"
> names(answer) <- first

which obviously is an error. The help is not explicit what should
happen when the length "of the first ... argument" is zero, but the
above behavior effectively does something like:

> answer <- list()
> first <- "A"
> names(answer) <- first[seq_along(answer)]
> answer
named list()

Is there a need for the docs to be updated, or should the result be an
unnamed empty list?

/Henrik

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel


Re: [Rd] mapply(): Special case of USE.NAMES=TRUE with recent R-devel updates

2021-11-30 Thread Hervé Pagès

And also:

  > mapply(paste, c(a="A"), character(), USE.NAMES = TRUE)

  Error in names(answer) <- names1 :

'names' attribute [1] must be the same length as the vector [0]


When the shortest arguments get recycled to the length of the longest, 
shouldn't their names also get recycled?


  > mapply(paste, c(a="A", b="B"), letters[1:6], USE.NAMES=TRUE)

  a b

  "A a" "B b" "A c" "B d" "A e" "B f"

That's assuming that rep() accurately materializes recycling (I hope it 
does):


  > rep(c(a="A", b="B"), length.out=6)

a   b   a   b   a   b

  "A" "B" "A" "B" "A" "B"


  > rep(c(a="A", b="B"), length.out=0)

  named character(0)


I always wished that the process of recycling which happens everywhere 
all the time in R was implemented in its own dedicated function 
recycle(). But that's another story.


Anyways, back to mapply(): Once what happens to the names during 
recycling is clarified, there should be no need to be explicit about 
what should happen when the length "of the first ... argument" is zero 
because it will no longer be a special case.


Cheers,
H.


On 30/11/2021 22:10, Henrik Bengtsson wrote:

Hi,

in R-devel (4.2.0), we now get:


mapply(paste, "A", character(), USE.NAMES = TRUE)

named list()

Now, in ?mapply we have:

USE.NAMES: logical; use the names of the first ... argument, or if
that is an unnamed character vector, use that vector as the names.

This basically says we should get:


answer <- list()
first <- "A"
names(answer) <- first


which obviously is an error. The help is not explicit what should
happen when the length "of the first ... argument" is zero, but the
above behavior effectively does something like:


answer <- list()
first <- "A"
names(answer) <- first[seq_along(answer)]
answer

named list()

Is there a need for the docs to be updated, or should the result be an
unnamed empty list?

/Henrik

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel



--
Hervé Pagès

Bioconductor Core Team
hpages.on.git...@gmail.com

__
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel