[R] How to modify object's code living in some environment?
Grzegorz Smoliński
g@@mo||n@k|1 @end|ng |rom gm@||@com
Mon Dec 27 15:36:01 CET 2021
Thank you, Ivan and Duncan. The truth is I still feel lost in this
concept of environments, generally speaking.
What I really thought about is the possibility to make a shiny app
where there will be possible to choose the function name (from the set
of all functions defined in this app / modules / other helper files
sourced on the start of app) and write on the beginning of the body of
this chosen function 'browser()'. I have developed an app using
modules and I have found that it is not convenient to stop the app,
write 'browser()' where needed, start the app and make something to
run the function where is `browser()`, so I have started to think if I
could write this `browser()` on demand, using the app itself (so:
without stopping the app), then make something to run the function,
where is 'browser()' and enabling interactive debugger. But I see now
it is not good to make too many assumptions didn't knowing some topic
:)
pon., 27 gru 2021 o 14:27 Duncan Murdoch <murdoch.duncan using gmail.com> napisał(a):
>
> On 27/12/2021 8:25 a.m., Duncan Murdoch wrote:
> > On 27/12/2021 8:06 a.m., Grzegorz Smoliński wrote:
> >> Hi,
> >>
> >> I know it is possible to find the environment in which some object
> >> lives using the 'environment()' function and the name of this object,
> >> but how to modify code of this object after this? Below is MRE:
> >
> > You are misunderstanding the relation between environments and
> > functions. However, your understanding is also being messed up by a bug
> > in R, so it's not entirely your fault.
> >
> > More details below:
> >>
> >> test <- function() 1
> >>
> >> test() # 1
> >>
> >> environment(test)$test <- eval(parse(text = "function() 2"))
> >
> > First: while it's usually true that function f will be found in
> > environment(f), these are really separate concepts.
> >
> > The environment of a function is where it will start looking for
> > non-local variables. It is initially the environment in which the
> > function definition is evaluated.
> >
> > The environment holding a function can be completely different. This
> > most commonly shows up when you have functions that create other
> > functions. For example,
> >
> > factory <- function(x) {
> > force(x)
> > function() print(x)
> > }
> >
> > f <- factory(3)
> > f()
> >
> > which will print 3. Here environment(f) is the evaluation frame of the
> > call to factory(3), while f is not stored there, it is stored in the
> > global environment. (If you run ls(environment(f)) you won't see "f"
> > listed, only "x".)
> >
> > In your first line "test <- function() 1" you create a function named
> > test in the global environment whose environment is also the global
> > environment.
> >
> > The third line should do what you want, i.e. put the new definition into
> > the environment of the test function, using the name test, but it
> > doesn't: I think that's a bug in R. It should be equivalent to
> >
> > .GlobalEnv$test <- eval(parse(text = "function() 2"))
> >
> > or
> >
> > test <- eval(parse(text = "function() 2"))
> >
> > but as you saw, it's not. They do what you were expecting your third
> > line to do.
> >
> >>
> >> test() # still 1
> >>
> >> .GlobalEnv$test <- eval(parse(text = "function() 3"))
> >>
> >> test() # 3
> >>
> >> The context is I have shiny app which code I would like to modify (add
> >> something) and the only way to refer to this environment I know is to
> >> use 'environment()' (as I see, those functions do not live in
> >> .GlobalEnv), but as you can see above, I can't use it to modify the
> >> code.
> >
> > You can, as long as you work around the bug in R. Just change that
> > third line into two statements:
> >
> > e <- environment(test)
> > e$test <- eval(parse(text = "function() 2"))
> >
>
> And I forgot to mention, as Ivan did, that this might not do what you
> want, because of the difference in the concepts of "environment(f)" and
> "environment where f is found".
>
> Duncan Murdoch
More information about the R-help
mailing list