R help - Oct 2001 - Assignment of structures on a given environment

Hi,

In order to avoid deep copies by passing large arguments to functions or
returning values, I'm trying to do the assignment of variables in a
given environment. The problem is when I try to assign a structure: a
list for example.

If I have:
    ind <- c("a","b")

my idea is doing something like

    l <- alist()
    l[ind] <- as.list(c(20,40))

 in a given environment.



Example:
    ref <- new.env()
    (.....)
    assign("l",alist(),env=ref)


    If I do
        assign("l$a",20,env=ref)

    it creates me a new variable in the ref environment  named "l$a"



    So,  I did:

    eval(l <- alist(), env=ref)

    but this creates the l list both on the current and on the ref
environment.


    The alternative solution that I found out was:

    evalq(l<-alist(),env=ref)

    and then

        evalq(ind <-c("a","b"), env=ref)
        evalq(l[ind] <- as.list(c(20,40)), env=ref)


I would like to know if there is another possible solution, instead of
doing these 'evalqs' along the program code.


Thanks,

    Rita


-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Hi,

In order to avoid deep copies by passing large arguments to functions or

returning values, I'm trying to do the assignment of variables in a
given environment. The problem is when I try to assign a structure: a
list for example.

If I have:
    ind <- c("a","b")

my idea is doing something like

    l <- alist()
    l[ind] <- as.list(c(20,40))

 in a given environment.



Example:
    ref <- new.env()
    (.....)
    assign("l",alist(),env=ref)


    If I do
        assign("l$a",20,env=ref)

    it creates me a new variable in the ref environment  named "l$a"



    So,  I did:

    eval(l <- alist(), env=ref)

    but this creates the l list both on the current and on the ref
environment.


    The alternative solution that I found out was:

    evalq(l<-alist(),env=ref)

    and then

        evalq(ind <-c("a","b"), env=ref)
        evalq(l[ind] <- as.list(c(20,40)), env=ref)


I would like to know if there is another possible solution, instead of
doing these 'evalqs' along the program code.


Thanks,

    Rita



-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

I was having trouble with a similar issue recently, until I was informed
that arguments to functions aren't copied unless they're altered. So I
can
pass a large matrix through a routine with as many sublevels as i need and
the matrix won't be recreated everytime I pass it to another function. Is
this the problem you're trying to avoid?

dave


On Tue, 16 Oct 2001, Rita Ribeiro wrote:
> Hi,
> 
> In order to avoid deep copies by passing large arguments to functions or
> returning values, I'm trying to do the assignment of variables in a
> given environment. The problem is when I try to assign a structure: a
> list for example.
> 
> If I have:
>     ind <- c("a","b")
> 
> my idea is doing something like
> 
>     l <- alist()
>     l[ind] <- as.list(c(20,40))
> 
>  in a given environment.
> 
> 
> 
> Example:
>     ref <- new.env()
>     (.....)
>     assign("l",alist(),env=ref)
> 
> 
>     If I do
>         assign("l$a",20,env=ref)
> 
>     it creates me a new variable in the ref environment  named
"l$a"
> 
> 
> 
>     So,  I did:
> 
>     eval(l <- alist(), env=ref)
> 
>     but this creates the l list both on the current and on the ref
> environment.
> 
> 
>     The alternative solution that I found out was:
> 
>     evalq(l<-alist(),env=ref)
> 
>     and then
> 
>         evalq(ind <-c("a","b"), env=ref)
>         evalq(l[ind] <- as.list(c(20,40)), env=ref)
> 
> 
> I would like to know if there is another possible solution, instead of
> doing these 'evalqs' along the program code.
> 
> 
> Thanks,
> 
>     Rita
> 
> 
>
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> 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
>
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
> 
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
|David Richmond                It works on a          |      
+ Dept. of Sociology          complex scientific      + 
|Saint Mary's College          principle, known as    |  
+ Notre Dame, IN 46556               "pot luck."      +
|219-284-4517                    - The Doctor         |
+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+

-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Hi,

I've already read something about it, and I known that when passing
arguments to
functions, R only passes a "promise" . The argument is copied only if
its value is
changed. But the problem is that I'm sucessively building a list along
the functions
calls, so the value is always updated....

Rita


David A Richmond wrote:
> I was having trouble with a similar issue recently, until I was
informed
> that arguments to functions aren't copied unless they're altered.
So I
can
> pass a large matrix through a routine with as many sublevels as i need
and
> the matrix won't be recreated everytime I pass it to another function.
Is
> this the problem you're trying to avoid?
>
> dave
>
> On Tue, 16 Oct 2001, Rita Ribeiro wrote:
>
> > Hi,
> >
> > In order to avoid deep copies by passing large arguments to
functions or
> > returning values, I'm trying to do the assignment of variables in
a
> > given environment. The problem is when I try to assign a structure:
a
> > list for example.
> >
> > If I have:
> >     ind <- c("a","b")
> >
> > my idea is doing something like
> >
> >     l <- alist()
> >     l[ind] <- as.list(c(20,40))
> >
> >  in a given environment.
> >
> >
> >
> > Example:
> >     ref <- new.env()
> >     (.....)
> >     assign("l",alist(),env=ref)
> >
> >
> >     If I do
> >         assign("l$a",20,env=ref)
> >
> >     it creates me a new variable in the ref environment  named
"l$a"
> >
> >
> >
> >     So,  I did:
> >
> >     eval(l <- alist(), env=ref)
> >
> >     but this creates the l list both on the current and on the ref
> > environment.
> >
> >
> >     The alternative solution that I found out was:
> >
> >     evalq(l<-alist(),env=ref)
> >
> >     and then
> >
> >         evalq(ind <-c("a","b"), env=ref)
> >         evalq(l[ind] <- as.list(c(20,40)), env=ref)
> >
> >
> > I would like to know if there is another possible solution, instead
of
> > doing these 'evalqs' along the program code.
> >
> >
> > Thanks,
> >
> >     Rita
> >
> >
> >
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> > 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
> >
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._
> >
>
> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
> |David Richmond                It works on a          |
> + Dept. of Sociology          complex scientific      +
> |Saint Mary's College          principle, known as    |
> + Notre Dame, IN 46556               "pot luck."      +
> |219-284-4517                    - The Doctor         |
> +-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+-+
>
>
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
> 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
>
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._




-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._

Hi,


Robert Gentleman wrote:
> Yes there is a way without evalq.
> Simply get and retrieve the list. From your description you say that
> you need to
>   1) keep the list structure
>   2) keep successively updating the list structure
>
> >From those perspectives putting the list in an environment will not
> help you save copies. In R we generally copy on mutate (not always and
> it could be improved but mostly).
>
> You really have not described the problem that you are trying to
> solve, rather you are giving us an abstraction and from that it is
>
The list I'm building keeps a variety of information about some dataset!
Among other things the list keeps the dataset itself in the data frame R format.

As there are some huge datasets, I was wishing to avoid frequent copies of that
structure.

I've already pass that list as an argument and there seems to be no problem
in
the execution. The only issue is that in very large datasets those frequent
copies could slow down the execution.

I've also stored this list in the Global environment, but as the program
produces also an output file with similar information I wouldn't like the
idea
of leaving some "garbage" in the Global environment or making users
manage some
global variable...... Ok, I can do 'rm', but if users do want to have
access to
the produced list??
My idea was making the program  return some kind of pointer to the environment
where is the list at, which users would use if they want to access it.... I
thought that way it wouldn't be necessary any copies....

Well, I'm seeing that there is no much way around it....


    Thanks  again for the tips,

    Rita


> very hard for anyone to give you sensible advice.
>
> I still truely doubt that size is an issue. Have you tried passing the
> list directly? Do you know that size is an issue or do you just
> presume it?
>
> Finally a very ugly solution that involves no evalq's
>
> You can always store your list in the Global environment and access it
> using get, assign and <<-. There should be only one copy but if you
> keep changing it, it will be duplicated, that's kind of necessary.
>
> --
>
+---------------------------------------------------------------------------+
> | Robert Gentleman                 phone : (617) 632-5250                  
|
> | Associate Professor              fax:   (617)  632-2444                  
|
> | Department of Biostatistics      office: M1B28
> | Harvard School of Public Health  email: rgentlem at
jimmy.dfci.harvard.edu   |
>
+---------------------------------------------------------------------------+
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
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
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._