Learn R Programming

rlang (version 0.2.2)

with_env: Evaluate an expression within a given environment

Description

These functions evaluate expr within a given environment (env for with_env(), or the child of the current environment for locally). They rely on eval_bare() which features a lighter evaluation mechanism than base R base::eval(), and which also has some subtle implications when evaluting stack sensitive functions (see help for eval_bare()).

Usage

with_env(env, expr)

locally(expr)

Arguments

env

An environment within which to evaluate expr. Can be an object with an get_env() method.

expr

An expression to evaluate.

Life cycle

These functions are experimental. Expect API changes.

Details

locally() is equivalent to the base function base::local() but it produces a much cleaner evaluation stack, and has stack-consistent semantics. It is thus more suited for experimenting with the R language.

Examples

Run this code
# NOT RUN {
# with_env() is handy to create formulas with a given environment:
env <- child_env("rlang")
f <- with_env(env, ~new_formula())
identical(f_env(f), env)

# Or functions with a given enclosure:
fn <- with_env(env, function() NULL)
identical(get_env(fn), env)


# Unlike eval() it doesn't create duplicates on the evaluation
# stack. You can thus use it e.g. to create non-local returns:
fn <- function() {
  g(get_env())
  "normal return"
}
g <- function(env) {
  with_env(env, return("early return"))
}
fn()


# Since env is passed to as_environment(), it can be any object with an
# as_environment() method. For strings, the pkg_env() is returned:
with_env("base", ~mtcars)

# This can be handy to put dictionaries in scope:
with_env(mtcars, cyl)
# }

Run the code above in your browser using DataLab