[R] Sequences

[Bert Gunter]

Aren't you mixing up [[ and [ ?

> x <- 1:3
> x[0]
integer(0)
> x[[0]]
Error in x[[0]] : attempt to select less than one element
> x[4]
[1] NA
> x[[4]]
Error in x[[4]] : subscript out of bounds

The docs say: 

"The default methods work somewhat differently for atomic vectors,
matrices/arrays and for recursive (list-like, see is.recursive) objects."

-- which we certainly see in the above differences. So even though both will
do the same thing in certain circumstances (e.g. x[1] = x[[1]] = 1), [[
should really be reserved for recursive (list-like) objects. Unfortunately,
there are certain circumstances where you need to use [ on lists: for
example, note that whether component names are preserved depends on how you
do things:

> y1 <- list(a=1:3)
> y2 <- list(b=4,d=5)
> c(y1,y2)
$a
[1] 1 2 3

$b
[1] 4

$d
[1] 5


> c(y1,y2[1])
$a
[1] 1 2 3

$b
[1] 4

> c(y1,y2[[1]])
$a
[1] 1 2 3

[[2]]
[1] 4


I think the detailed  "explanation" of this behavior is that [[ selects a
component, the atomic vector with single element 4,  while [ selects a list
containing this atomic vector as it's first component with name "b". c()
then does it's thing: in the first 2 cases, it just concatenates the two
lists; in the third, it first creates a list containing the single unnamed
atomic vector and then concatenates this with y1. Clarification/correction
of this would be appreciated if I haven't got it right or am missing
something important.

I've found V&R's S PROGRAMMING helpful in clarifying some of these semantic
complexities (and sometimes inconsistencies?), but I would appreciate any
suggestions for other and especially more complete explanations of all of
this, as V&R doesn't cover everything (and is more S language focused rather
than R-centric in detail).  I think it's fair to say that the man pages do
not provide all of the details, either. As a result, I still occasionally
get bitten by these indexing subtleties (which, of course, may just be due
to my dummheit).

Cheers,
Bert


Bert Gunter
Genentech Nonclinical Biostatistics
650-467-7374


-----Original Message-----
From: r-help-bounces at r-project.org [mailto:r-help-bounces at r-project.org] On
Behalf Of Stavros Macrakis
Sent: Tuesday, April 07, 2009 9:11 AM
To: Melissa2k9
Cc: r-help at r-project.org
Subject: Re: [R] Sequences

On Tue, Apr 7, 2009 at 8:13 AM, Melissa2k9
<m.mcquillan at lancaster.ac.uk>wrote:

>
> I am trying to make a sequence and am  using a for loop for this. I want
to
> start off with an initial value ie S[0]=0 then use the loop to create
other
> values. This is what I have so far but I just keep getting error messages.
>

R only allows positive integer subscripts, and defines s[0] as the empty
vector.

You might think that assigning to an empty vector would give an error, but R
semantics say instead that this is a no-op. This is so that things like s[
<<some boolean condition>> ] <- value will assign the value to all elements
of s meeting the condition -- while allowing for the possibility that none
of them meet that condition.

If you try to assign to an illegal individual element, however, you do get
an error:

     s[[0]] <- 45       # gives an error

R's subscripting semantics are very convenient for many operations, but do
take some getting used to.

            -s

	[[alternative HTML version deleted]]

______________________________________________
R-help at r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.