changing default environment for assignment of new variables

Question

I often want to create many variables in an environment under the global environment. This can be done easily with the envir argument to sys.source -- if all of the variables created by the file that one is sourcing are supposed to go into a single environment.

But I typically work with a file that creates sets of variables. One set should go into one environment, another set should go into another environment, and so on. I don't want to split this file into multiple files and then make multiple calls to sys.source.

Instead, I would like a command that lets me change the default environment for assignment of new variables. For example:

e <- new.env()
setDefaultEnvironment(e)
tmp <- 2
e$tmp           #  2 
.GlobalEnv$tmp  #  Error: object 'tmp' not found

But setDefaultEnvironment isn't a real command.

Is there any safe way to do this sort of thing in R?

Answer 1

Perhaps someone can improve on this, removing the need to quote the variable name:

e <- new.env()
`%=%` <- function(x,y) {assign(x,y,envir = e)}

"d" %=% 5

e$d
[1] 5

But this feels kind of silly to me. Maybe just use assign directly? More typing, maybe, but it does what you want with less danger.

Answer 2

The evalq function will evaluate its first argument in a specified environment, so you could create your new environment, then wrap the assignments into evalq.

Answer 3

Instead changing the environment, an easy way to do this is to save all variables in a list.

e <- list()
e$tmp <- 2