R help - Oct 2013 - How to pass an object between two independant environment?

Hello everyone,

I am trying to pass an object cleanly between two independant 
environments, any ideas on how to do that?

I have something like this:

calling.function <- function(my.object) {
... ...
local(source("function_to_which_I_want_to_pass_my_object.Rda", local =
new.env(parent = .GlobalEnv))) # I can't change this part of the code
}

The calling function will be part of a package, so I don't want to use 
the global environment (that's the user's space, not a space that should
be used by the packages). Writing a temp file is unclean for the same 
reason (and using the user's hard-drive is actually worse than using its 
space in R!). The package namespace is locked when the package is 
loaded, so using the package environment is not a perfect option either.

I was wondering, is there a way to create a named environment in R? If 
it was possible to do that, then the calling function could 'assign' to 
the named-environment and the scripted function could 'get' from it and 
the problem would be solved.

Thanks for your help,

Frederic

On 02/10/2013 10:12 AM, Frederic Fournier wrote:
> Hello everyone,
>
> I am trying to pass an object cleanly between two independant
> environments, any ideas on how to do that?
>
> I have something like this:
>
> calling.function <- function(my.object) {
> ... ...
> local(source("function_to_which_I_want_to_pass_my_object.Rda",
local > new.env(parent = .GlobalEnv))) # I can't change this part of the
code
> }
>
> The calling function will be part of a package, so I don't want to use
> the global environment (that's the user's space, not a space that
should
> be used by the packages). Writing a temp file is unclean for the same
> reason (and using the user's hard-drive is actually worse than using
its
> space in R!). The package namespace is locked when the package is
> loaded, so using the package environment is not a perfect option either.
>
> I was wondering, is there a way to create a named environment in R? If
> it was possible to do that, then the calling function could
'assign' to
> the named-environment and the scripted function could 'get' from it
and
> the problem would be solved.
name <- new.env()

creates an environment named "name".

Duncan Murdoch