[R] How to set a former environment?

[Duncan Murdoch]

On 4/3/2006 1:44 PM, Werner Wernersen wrote:
> Hi,
> 
> I probably misunderstand the entire concept of
> environments in R. I now have a list of environments
> in which each has a number of variables. What I want
> to do is to apply a function on each of those
> environments which adds additional variables to that
> particular environment. Is there a way to set the
> environment of the function so that also the variables
> newly created in the function are added to the
> environment?
> 
> Right now I do
> environment(FN) <- oldEnvironment
> before I call the function and the function operates
> on the old environments variables but the newly
> created ones are not in the old environment.

The line

  environment(FN) <- oldEnvironment

sets the closure of the function, i.e. the parent of the evaluation 
environment.  Assignments within the function are made within the 
evaluation environment, which usually disappears when the function 
returns.  (An exception to this:  if you create and return a function 
within a function, the closure of the new one is by default the 
evaluation environment of the old one, so it will outlast the outer 
function call.)

Generally there is no way to set the evaluation environment of a 
function, but you could write a function that copied things from it 
somewhere else, e.g.

copyObjects <- function(fromEnvironment, toEnvironment) {
   names <- ls(fromEnvironment)
   for (n in names) assign(n, get(n, env=fromEnvironment), 
env=toEnvironment)
}

and then putting

  copyObjects(environment(), parent.env(environment()))

in a function will copy all the locals into the parent.  More simply, 
you can do assignments directly to the parent environment using

  assign("name", env=parent.env(environment()))