R help - Dec 2022 - difference between script and a function....

Dear Simmons,
                            Thanks a lot. One more question:


  1.  If there is  some error in the script, the error will be output to the
stdout? Or to the file that it creates for saving the output of the script?

Thanking you,
Yours sincerely,
AKSHAY M KULKARNI
________________________________
From: Andrew Simmons <akwsimmo at gmail.com>
Sent: Saturday, December 24, 2022 8:18 PM
To: akshay kulkarni <akshay_e4 at hotmail.com>
Cc: R help Mailing list <r-help at r-project.org>
Subject: Re: [R] difference between script and a function....

1. The execution environment for a script is the global environment. Each R
script run from a shell will be given its own global environment. Each R session
has exactly one global environment, but you can have several active R sessions.

2. Using return in a script instead of a function will throw an error

Error: no function to return from, jumping to top level

3. You can use print(), cat(), and writeLines() to write to the output. It does
not save your R objects before it exits the script. You could use save() or
save.image() to save your objects, or possibly saveRDS() if you are only looking
to save one object. You could also use source() if you just want the objects
from another script.

4. Will you have shared access to the objects in another R session? No, objects
are not shared, unless you've got something weird setup with external
pointers. Each session has it owns global environment.

5. Any of the doc pages for the functions I listed above would help, you can
also do

?utils::Rscript

	[[alternative HTML version deleted]]

On Sat, 24 Dec 2022 14:54:52 +0000
akshay kulkarni <akshay_e4 at hotmail.com> wrote:
> If there is  some error in the script, the error will be output to
> the stdout? Or to the file that it creates for saving the output of
> the script?
When using Rscript: to stderr, to be precise.

When using R CMD BATCH: to the Rout file.
> Will ALL the intermediate objects be stored ?
May I shamelessly plug my own package, depcache? Using its cache()
function, you can save intermediate objects into *.rds files in a
relatively transparent manner. By default, they are stored in a
subdirectory of the working directory. A downside is that I haven't
come up with a useful cache expiration strategy yet.

A more involved approach that will also let you access your
intermediates is provided by the "targets" package, not only on CRAN
but also peer-reviewed by rOpenSci.

-- 
Best regards,
Ivan