Weird feature when creating function lists with apply

Hi R-fellows,

can anyone explain this weird feature? I am trying to create a list of
functions with apply, and it appears that there is something strange going
on when the function evaluates the argument a. When I duplicate a into a
local variable b, the result changes !?!

Any pointers? Thank in advance. Cheers,

Uffe


# Create a function which returns a function

f1 <- function(a) {return(function(x) a*x)}

# Create a list of functions, parameterised by 1:4

apply(as.array(1:4),1,f1)[[2]](1)

[1] 4

# Repeat

f2 <- function(a) {b <- a ;return(function(x) b*x)}
apply(as.array(1:4),1,f2)[[2]](1)

[1] 2

I don't think this is what you intended.

ll <- apply(as.array(1:4),1,f1)
ll
[[1]]
function(x) a*x
<environment: a6d374>

[[2]]
function(x) a*x
<environment: a6d09c>

[[3]]
function(x) a*x
<environment: a6dd60>

[[4]]
function(x) a*x
<environment: a6da88>

which is not what your comment implies.

And

get("a", envir=environment(ll[[2]]))

[1] 4

shows why they differ.


1) This is a strange way to produce a list via arrays:
ll <- lapply(1:4, f1) is what is normally used.

2) I think what you really wanted was

f3 <- function(a) substitute(function(x) a*x, list(a = a))
lapply(1:4, f3)
[[1]]
function(x) 1 * x

[[2]]
function(x) 2 * x

[[3]]
function(x) 3 * x

[[4]]
function(x) 4 * x

Brian D. Ripley,                  ripley at stats.ox.ac.uk
Professor of Applied Statistics,  http://www.stats.ox.ac.uk/~ripley/
University of Oxford,             Tel:  +44 1865 272861 (self)
1 South Parks Road,                     +44 1865 272860 (secr)
Oxford OX1 3TG, UK                Fax:  +44 1865 272595

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help 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-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Uffe H?gsbro Thygesen <uht at dfu.min.dk> writes:

# Create a function which returns a function

f1 <- function(a) {return(function(x) a*x)}

# Create a list of functions, parameterised by 1:4

apply(as.array(1:4),1,f1)[[2]](1)

[1] 4

# Repeat

f2 <- function(a) {b <- a ;return(function(x) b*x)}
apply(as.array(1:4),1,f2)[[2]](1)

[1] 2

This should work too:
f3 <- function(a) {a;function(x) a*x}
apply(as.array(1:4),1,f3)[[2]](1)

It is essentially the same situation that caused Luke Tierney
substantial headscratching recently. The key is lazy evaluation.
Here's part of the conversation from then:

[Snip from email from Luke]

Luke Tierney <luke at nokomis.stat.umn.edu> writes:

mkf <- function(j) function() j
for (i in 1:10) x[[i]] <- mkf(i)

That doesn't work either:

x[[2]]()

[1] 10

but for a slightly different reason: Lazy evaluation defers evaluating
the argument to mkf(i) until the funtction is called.

Argh. I've been looking at this for ten minutes and it still refuses
to sink in....

Now imagine getting bitten by this while experimenting with some new
threading code, which is of course then the prime suspect for all bad
things that happen, and you can see why Duncan and I were tearing our
hair out for a bit :-).

Oh, now I see it:

z<-2 ; f<-mkf(z)
z<-3
f()

[1] 3

z<-4
f()

[1] 3

The return value of mkf is not really a function of the argument, it
is the same function every time, namely the function returning "j"
which it will look up in its parent environment (the lexical one...)
when called, and only *then* will it force the promise. 

O__  ---- Peter Dalgaard             Blegdamsvej 3  
  c/ /'_ --- Dept. of Biostatistics     2200 Cph. N   
 (*) \(*) -- University of Copenhagen   Denmark      Ph: (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)             FAX: (+45) 35327907
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help 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-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Hello All,

I am creating a graph in a loop, and want to use letters for marking
symbols.  Is there a function to convert an integer into a character.  I
am looking for something similar to the ASC() function of good old
BASIC, or the (char) type conversion of C.

Thanks,

John.
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help 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-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

LETTERS or letters might do what you want; they are character vectors
containing the (upper- or lowercase) alphabet.  (If they didn't exist you
could create them ...)

Hello All,

I am creating a graph in a loop, and want to use letters for marking
symbols.  Is there a function to convert an integer into a character.  I
am looking for something similar to the ASC() function of good old
BASIC, or the (char) type conversion of C.

Thanks,

John.
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help 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-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

318 Carr Hall                                bolker at zoo.ufl.edu
Zoology Department, University of Florida    http://www.zoo.ufl.edu/bolker
Box 118525                                   (ph)  352-392-5697
Gainesville, FL 32611-8525                   (fax) 352-392-3704

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help 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-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

On Wed, 10 Oct 2001 09:55:25 -0300, you wrote in message
<3BC4453D.94ACB0E0 at acadiau.ca>:

Hello All,

I am creating a graph in a loop, and want to use letters for marking
symbols.  Is there a function to convert an integer into a character.  I
am looking for something similar to the ASC() function of good old
BASIC, or the (char) type conversion of C.

There are the letters and LETTERS vectors.  For example,

 > letters[1:3]
 [1] "a" "b" "c"

I don't think there's any function that does general ASCII
conversions, but you could fairly easily enter your own vector of
characters if letters aren't sufficient.  I imagine you'll have a hard
time entering a null character.  If you're on some platforms, you
could write your own general typecasting function using a fifo, but
those aren't supported everywhere.

Duncan Murdoch
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help 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-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._