Re: indexing feature (WAS indexing within a list using Splus

[Patrick Burns <pburns@pburns.seanet.com>]

Here is a specific example of what nerdy ole' Bill is talking about.

When making plots for publications, I like to write functions that create
the plots.  This means that months or years later there is no mucking about
trying to find how precisely that plot was made -- the data are stored inside
the function along with the plot commands.

The first step is to create a template function:

P.template <- function(file="xxx.eps")
{
        x <- 234
        if(length(file)) {
                # set up hardcopy plotting commands
        } else {
                # set up plotting to a window
        }
        plot(x)
}

Then create a specific instance:

P.description <- P.template

Then  specialize the instance:

P.description[[c(2,1,2)]] # make sure I have the right spot, should be 234
P.description[[c(2,1,2)]] <- the.real.x.object
P.description(NULL)  # look to see how well it works
P.description() # creates the file, the name in the function should be edited


In R, functions are not visibly lists like this (as well as R not having
this form
of subscripting).  The same thing can be done in R of course (it is still
S after 
all), but not as neatly to my knowledge.

Patrick Burns
pburns@pburns.seanet.com

http://www.seanet.com/~pburns/Spoetry
http://www.i-way.co.uk/~pat/Spoetry


Bill.Venables@CMIS.CSIRO.AU wrote:
> Just a minor footnote to Brian's reply below.  The only place where this
> kind of indexing does seem to be useful is in dealing with language objects.
> lists of lists of lists ... are not all that useful, or usual, but in
> language elements expressions of calls of calls of .... are the norm.
>
> Dealing with language objects does indeed take a 'certain type' of mindset -
> the kind that revels in programming puzzles and challenges.  Nerds like me,
> in other words...:-)
>
> Bill Venables.
>
> -----Original Message-----
> From: ripley@stats.ox.ac.uk [mailto:ripley@stats.ox.ac.uk]
> Sent: Monday, October 28, 2002 5:45 PM
> To: Nick.Ellis@csiro.au
> Cc: s-news@wubios.wustl.edu
> Subject: Re: [S] indexing feature (WAS indexing within a list using
> Splus 
>
>
> I think that is cart-before-horse.  It is not implemented in R, and there
> was a request (nay, a bug report, PR#1558, by one of your colleagues) on
> this.  It just does not occur in any of the commonly-used S library
> sections which have been ported to R.
>
> I surmised at that time that people have not found it useful.  I never
> have.  It must take a certain type of problem (or mindset) to work with
> lists of lists of lists.
>
> Brian
>
> On Mon, 28 Oct 2002 Nick.Ellis@csiro.au wrote:
>
> > I got one reply on this indexing feature from Bill Venables. He pointed
> >     
> out
> > that this form of indexing does not occur in R. That may explain why it is
> > rarely used, as one would try to avoid incompatibilities between the two
> > flavours of S.
> >
> > Splus> list(list(list(1)))[[c(1,1,1)]]
> > [1] 1
> >
> > R> list(list(list(1)))[[c(1,1,1)]]
> > Error: attempt to select more than one element
> >
> > Nick Ellis
> > CSIRO Marine Research	mailto:Nick.Ellis@csiro.au
> > PO Box 120			ph    +61 (07) 3826 7260
> > Cleveland QLD 4163    	fax   +61 (07) 3826 7222
> > Australia			http://www.marine.csiro.au
> >
> > > -----Original Message-----
> > > From: Nick.Ellis@csiro.au [mailto:Nick.Ellis@csiro.au]
> > > Sent: Thursday, October 24, 2002 10:49 AM
> > > To: s.su@qut.edu.au
> > > Cc: s-news@lists.biostat.wustl.edu
> > > Subject: Re: [S] indexing within a list using Splus 2000 and Splus 6
> > >
> > > For instance
> > > A[[c(1,2,3)]] is the same
> > > as A[[1]][[2]][[3]].
> > >
> > > To the s-news community:
> > > I don't remember anyone using this form of indexing on
> > > s-news. I wonder if
> > > anyone routinely uses this kind of indexing, and would like
> > > to share any
> > > tricks? The feature is not described under help("[["); you
> > > have to go to the
> > > Programmers' Guide, Chapter 4, section "Subscripting Li
> > > sts".
> > >         
> >         
>
>