replicating lists

Yes, I agree that the work-around needs some care, but I'm not sure 
we need to wrap the list itself -- we're simply using properties of 
"[" on lists.  If the subscript is rep(1, 10), then we'll get the 
first element 10 times (which is *not* replicating the list).
To get the replication we need to index every element in 
the list as many times as required.

tmp <- array(fred, c(3, 4))

tmp <- array(list(fred)[rep(1, 12)], c(3, 4))

tmp[[3, 4]] # fred again!

hi,

Yes, I agree that the work-around needs some care, but I'm not sure
we need to wrap the list itself -- we're simply using properties of
"[" on lists.  If the subscript is rep(1, 10), then we'll get the
first element 10 times (which is *not* replicating the list).
To get the replication we need to index every element in
the list as many times as required.

David

Date: Wed, 11 Apr 2001 15:28:02 +0100 (BST)
From: Jonathan Rougier <J.C.Rougier@durham.ac.uk>
X-Sender: dma0jcr@laplace
To: David James <dj@research.bell-labs.com>
cc: r-devel@stat.math.ethz.ch
Subject: Re: [Rd] replicating lists
MIME-Version: 1.0

On Wed, 11 Apr 2001, David James wrote:

A workaround is to use rep() inside "[":

a <- list(a=1, b=2, c=3)
a[rep(1:3, 1:3)]

$a
[1] 1

$b
[1] 2

$b
[1] 2

$c
[1] 3

$c
[1] 3

$c
[1] 3

Note, however, that names are no longer unique!

This work-around is cunning but needs a bit of care, as to get the
replication of the whole of fred you need to wrap the list itself:

fred <- list(happy = 1:10, fred = "squash")
list(fred)[rep(1, 10)] # list with 10 copies of fred

Cheers, Jonathn.

Jonathan Rougier                       Science Laboratories
Department of Mathematical Sciences    South Road
University of Durham                   Durham DH1 3LE
tel: +44 (0)191 374 2361, fax: +44 (0)191 374 7388
http://www.maths.dur.ac.uk/stats/people/jcr/jcr.html

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-devel mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-devel-request@stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

fred <- list(happy = 1:10, name = "squash")
list(fred)[rep(1, 5)]                          # version 1
fred[rep(seq(along=fred), length=5)]           # version 2
unlist(list(fred)[rep(1, 5)], recursive=FALSE) # version 3

tmp <- array(fred, c(3, 4))

The most obvious fix is internal, and that should be done fairly soon.


On Wed, 11 Apr 2001, David James wrote:

hi,

Yes, I agree that the work-around needs some care, but I'm not sure
we need to wrap the list itself -- we're simply using properties of
"[" on lists.  If the subscript is rep(1, 10), then we'll get the
first element 10 times (which is *not* replicating the list).
To get the replication we need to index every element in
the list as many times as required.

David

Date: Wed, 11 Apr 2001 15:28:02 +0100 (BST)
From: Jonathan Rougier <J.C.Rougier@durham.ac.uk>
X-Sender: dma0jcr@laplace
To: David James <dj@research.bell-labs.com>
cc: r-devel@stat.math.ethz.ch
Subject: Re: [Rd] replicating lists
MIME-Version: 1.0

On Wed, 11 Apr 2001, David James wrote:

A workaround is to use rep() inside "[":

a <- list(a=1, b=2, c=3)
a[rep(1:3, 1:3)]

$a
[1] 1

$b
[1] 2

$b
[1] 2

$c
[1] 3

$c
[1] 3

$c
[1] 3

Note, however, that names are no longer unique!

This work-around is cunning but needs a bit of care, as to get the
replication of the whole of fred you need to wrap the list itself:

fred <- list(happy = 1:10, fred = "squash")
list(fred)[rep(1, 10)] # list with 10 copies of fred

Cheers, Jonathn.

Hi Brian,

After a discussion with David James can you clarify exactly what you think
the behaviour of "rep" will be?  The three alternatives for rep(fred, 5)
would appear to be

fred <- list(happy = 1:10, name = "squash")
list(fred)[rep(1, 5)]                          # version 1
fred[rep(seq(along=fred), length=5)]           # version 2
unlist(list(fred)[rep(1, 5)], recursive=FALSE) # version 3

I was expecting the first, as this makes sense in the context of

tmp <- array(fred, c(3, 4))

but I can see that the second is also a natural way to think about
replicating a list.  I also notice that S+ uses version 3.  Personally I
don't like version 3 as it is going to duplicate names, and it is not
going to fit with "array".

Confused, Jonathan.

Prof Brian Ripley writes:

On Thu, 12 Apr 2001, Jonathan Rougier wrote:

Hi Brian,

After a discussion with David James can you clarify exactly what you think
the behaviour of "rep" will be?  The three alternatives for rep(fred, 5)
would appear to be

fred <- list(happy = 1:10, name = "squash")
list(fred)[rep(1, 5)]                          # version 1
fred[rep(seq(along=fred), length=5)]           # version 2
unlist(list(fred)[rep(1, 5)], recursive=FALSE) # version 3

I was expecting the first, as this makes sense in the context of

tmp <- array(fred, c(3, 4))

but I can see that the second is also a natural way to think about
replicating a list.  I also notice that S+ uses version 3.  Personally I
don't like version 3 as it is going to duplicate names, and it is not
going to fit with "array".

Confused, Jonathan.

We follow the prototype in cases of doubt, so version 3.  *However*,
S-PLUS does different things by version, so version 6.0 drops the
duplicated names.  I really think only 3 makes sense: a list is a
generic vector, and you do exactly the same thing as for any other
sort of vector.  The only issue is the names, and

x <- letters[1:3]
names(x) <- x
rep(x, 5)

drops names in S+6.0, keeps them in S+3.4 and R.  So the S-PLUS
versions are doing consistent things with character vectors and lists.