R help - Sep 2011 - How to use a variable after $ ($x) for subscripting a list, How to source from a character object (source(x))

Dear R colleagues,

as result of a function a huge list is generated. From this result list
I'd like to extract information.

Think of the list i.e. as an object named "listResult" with the
following form:

[[a]]
 [1] [2] [3] [4] [5]

[[b]]
 [1] [2] [3] [4] [5]

[[c]]
 [1] [2] [3] [4] [5]

where levels=c(a,b,c)

I'd like to extract a data.frame like

a [2]
b [2]
c [2]

What I tried, is a function like this:

library(foreach)
loopExtract <- function(input) {
foreach(i=1:length(levels)) %do% {listResult[[i]]$input}
...

where the name of the variable [2] is meant to be the input.

Unfortunately it turned out, that the "input"-variable after the $ is
not substituted as expected. Subscripting the list with a variable after
the $ seems not possible.
The only workaround I found for that is a function like

loopExtract <- function(input) {
codetemplate <- as.character("result <- foreach(i=1:length(levels))
%do%
{listResult[[i]]$input)}")
require(stringr)
write(str_replace_all(codetemplate, "input", input),
file="tmp.r")
source("tmp.r")
return(result)
}

in other words I stored a template of the desired code as characters in
an object, substituted the "input" string in that character object,
wrote the result to a file and sourced the code of that file.

I stored the result in a file cause the expression source(codetemplate)
did not work. From the documentation I learned, that one can only source
"connections". And there seems no chance to establish a connection to
an
object. (Probably no one else, except me, has such a strange idea.)

Well, it works that way, but even though I am not a real programmer, I
realize how dirty this solution is. There must be a much simpler and
elegant way.

To sum it all up, my problem can be reduced to the following two questions

(1) How to subscript a list with a variable after $
(2) How to source from an object containing a code template
or (probably the most promising) (3) How to avoid both ;)

Unfortunately I am not experienced enough to find the solution. Please help!

Greetings from sunny Munich, Felix

I had a little bit of trouble following what exactly you mean to do
(specifically, I'm having a little bit of trouble understanding your sample
object "listResult" -- is it a list or a set of factors or a list of
lists
or something else entirely), but I believe that replacing the `$` operator
with additional copies of  `[` or `[[` as appropriate should get the job
done. $ is, for almost all purposes, a less flexible version of `[[`.

Consider the following:

R> a = list(a1 = rnorm(5), rnorm(6), rnorm(7))
R> b = list(runif(5), b2 = runif(6), runif(7))
R> c = list(-runif(5), -runif(6), c3 = -runif(7))

R> A = list(a = a,b = b,c = c)

R> for( i in seq_along(A)) { print(A[[i]][[3]][i]) }

which seems to work for me (even if it is frustratingly opaque as to its
function). Specifically,

R> identical(A$a, A[["a"]])
TRUE

R> identical(A$a,A[[1]])
TRUE

If this doesn't work, please clarify what exactly the object you have is and
what you are trying to get out of it and I'll give a more concrete answer.

On Tue, Sep 13, 2011 at 4:11 AM, drflxms <drflxms@googlemail.com> wrote:
> Dear R colleagues,
>
> as result of a function a huge list is generated. From this result list
> I'd like to extract information.
>
> Think of the list i.e. as an object named "listResult" with the
> following form:
>
> [[a]]
>  [1] [2] [3] [4] [5]
>
> [[b]]
>  [1] [2] [3] [4] [5]
>
> [[c]]
>  [1] [2] [3] [4] [5]
>
> where levels=c(a,b,c)
>
> I'd like to extract a data.frame like
>
> a [2]
> b [2]
> c [2]
>
> What I tried, is a function like this:
>
> library(foreach)
> loopExtract <- function(input) {
> foreach(i=1:length(levels)) %do% {listResult[[i]]$input}
> ...
>
> where the name of the variable [2] is meant to be the input.
>
> Unfortunately it turned out, that the "input"-variable after the
$ is
> not substituted as expected. Subscripting the list with a variable after
> the $ seems not possible.
> The only workaround I found for that is a function like
>
> loopExtract <- function(input) {
> codetemplate <- as.character("result <-
foreach(i=1:length(levels)) %do%
> {listResult[[i]]$input)}")
> require(stringr)
> write(str_replace_all(codetemplate, "input", input),
file="tmp.r")
> source("tmp.r")
> return(result)
> }
>
> in other words I stored a template of the desired code as characters in
> an object, substituted the "input" string in that character
object,
> wrote the result to a file and sourced the code of that file.
>
> I stored the result in a file cause the expression source(codetemplate)
> did not work. From the documentation I learned, that one can only source
> "connections". And there seems no chance to establish a
connection to an
> object. (Probably no one else, except me, has such a strange idea.)
>
> Well, it works that way, but even though I am not a real programmer, I
> realize how dirty this solution is. There must be a much simpler and
> elegant way.
>
> To sum it all up, my problem can be reduced to the following two questions
>
> (1) How to subscript a list with a variable after $
> (2) How to source from an object containing a code template
> or (probably the most promising) (3) How to avoid both ;)
>
> Unfortunately I am not experienced enough to find the solution. Please
> help!
>
> Greetings from sunny Munich, Felix
>
> ______________________________________________
> R-help@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.
>
	[[alternative HTML version deleted]]

On Sep 13, 2011, at 4:11 AM, drflxms wrote:
> Dear R colleagues,
>
> as result of a function a huge list is generated. From this result  
> list
> I'd like to extract information.
>
> Think of the list i.e. as an object named "listResult" with the
> following form:
>
> [[a]]
> [1] [2] [3] [4] [5]
>
> [[b]]
> [1] [2] [3] [4] [5]
>
> [[c]]
> [1] [2] [3] [4] [5]
>
> where levels=c(a,b,c)
>
> I'd like to extract a data.frame like
>
> a [2]
> b [2]
> c [2]
>
Naming the last dog "cc" instead of "c":

 > as.data.frame(lapply(listResult, "[", 2) )
   a b cc
1 2 2  2


> What I tried, is a function like this:
>
> library(foreach)
> loopExtract <- function(input) {
> foreach(i=1:length(levels)) %do% {listResult[[i]]$input}
> ...
>
> where the name of the variable [2] is meant to be the input.
It would have been had you used this form:
> foreach(i=1:length(levels)) %do% {listResult[[i]][[input]]}
>
> Unfortunately it turned out, that the "input"-variable after the
$ is
> not substituted as expected. Subscripting the list with a variable  
> after
> the $ seems not possible.
Correct. That is the main difference between "$" and "[["
and the
reason "$" is far less useful in programming situations.
> The only workaround I found for that is a function like
>
> loopExtract <- function(input) {
> codetemplate <- as.character("result <-
foreach(i=1:length(levels))
> %do%
> {listResult[[i]]$input)}")
> require(stringr)
> write(str_replace_all(codetemplate, "input", input),
file="tmp.r")
> source("tmp.r")
> return(result)
> }
>
> in other words I stored a template of the desired code as characters  
> in
> an object, substituted the "input" string in that character
object,
> wrote the result to a file and sourced the code of that file.
>
> I stored the result in a file cause the expression  
> source(codetemplate)
> did not work. From the documentation I learned, that one can only  
> source
> "connections". And there seems no chance to establish a
connection
> to an
> object. (Probably no one else, except me, has such a strange idea.)
>
> Well, it works that way, but even though I am not a real programmer, I
> realize how dirty this solution is. There must be a much simpler and
> elegant way.
>
> To sum it all up, my problem can be reduced to the following two  
> questions
>
> (1) How to subscript a list with a variable after $
> (2) How to source from an object containing a code template
> or (probably the most promising)
I'm afraid you lost me after a while. If using "[[" with a
character-
valued argument which is intended to be interpreted is not what you  
needed,  then make a slightly less wordy posting that has an example  
of what you wanted.
> (3) How to avoid both ;)
Avoid .... what?
>
> Unfortunately I am not experienced enough to find the solution.  
> Please help!
>
> Greetings from sunny Munich, Felix
>
-- 

David Winsemius, MD
West Hartford, CT