Is there a way to identify all the workspace objects created, modified or referenced in a sourced script? I have hundreds of randomly-named objects in my workplace and am 'cleaning house' - I would like to be able to be more proactive about this in the future and just rm() the heck out of the sourced script at the end.
The simplest way is to store your environment objects in a list before sourcing, sourcing, then comparing the new environment objects with the old list.
Here is some pseudo-code.
old_objects <- ls()
source(file)
new_objects <- setdiff(ls(), c(old_objects, "old_objects"))
This will identify the created objects. To identify whether an object was modified, I don't see another solution than to store all your objects in a list beforehand and then running identical afterwards.
# rm(list = ls(all = TRUE))
a <- 1
b <- 1
old_obj_names <- ls()
old_objects <- lapply(old_obj_names, get)
names(old_objects) <- old_obj_names
# source should go here
a <- 2
c <- 3
# I add "old_obj_names" and "old_objects" in the setdiff as these
# were created after the call to ls but before the source
new_objects <- setdiff(ls(), c(old_obj_names, "old_obj_names", "old_objects"))
modified_objects <- sapply(old_obj_names, function(x) !identical(old_objects[[x]], get(x)),
USE.NAMES = TRUE)
modified_objects <- names(modified_objects[modified_objects])
new_objects is indeed "c" and modified_objects is indeed "a" in this example. Obviously, for this to work, you need to ensure that neither old_objects nor old_obj_names are in any way created or modified in the sourced file!