Learn R Programming

drake (version 7.13.10)

read_drake_seed: Read the pseudo-random number generator seed of the project. [Stable]

Description

When a project is created with make() or drake_config(), the project's pseudo-random number generator seed is cached. Then, unless the cache is destroyed, the seeds of all the targets will deterministically depend on this one central seed. That way, reproducibility is protected, even under randomness.

Usage

read_drake_seed(path = NULL, search = NULL, cache = NULL, verbose = NULL)

Value

An integer vector.

Arguments

path

Path to a drake cache (usually a hidden .drake/ folder) or NULL.

search

Deprecated.

cache

drake cache. See new_cache(). If supplied, path is ignored.

verbose

Deprecated on 2019-09-11.

Examples

Run this code
if (FALSE) {
isolate_example("contain side effects", {
cache <- storr::storr_environment() # Just for the examples.
my_plan <- drake_plan(
  target1 = sqrt(1234),
  target2 = sample.int(n = 12, size = 1) + target1
)
tmp <- sample.int(1) # Needed to get a .Random.seed, but not for drake.
digest::digest(.Random.seed) # Fingerprint of the current R session's seed.
make(my_plan, cache = cache) # Run the project, build the targets.
digest::digest(.Random.seed) # Your session's seed did not change.
# drake uses a hard-coded seed if you do not supply one.
read_drake_seed(cache = cache)
readd(target2, cache = cache) # Randomly-generated target data.
clean(target2, cache = cache) # Oops, I removed the data!
tmp <- sample.int(1) # Maybe the R session's seed also changed.
make(my_plan, cache = cache) # Rebuild target2.
# Same as before:
read_drake_seed(cache = cache)
readd(target2, cache = cache)
# You can also supply a seed.
# If your project already exists, it must agree with the project's
# preexisting seed (default: 0)
clean(target2, cache = cache)
make(my_plan, cache = cache, seed = 0)
read_drake_seed(cache = cache)
readd(target2, cache = cache)
# If you want to supply a different seed than 0,
# you need to destroy the cache and start over first.
clean(destroy = TRUE, cache = cache)
cache <- storr::storr_environment() # Just for the examples.
make(my_plan, cache = cache, seed = 1234)
read_drake_seed(cache = cache)
readd(target2, cache = cache)
})
}

Run the code above in your browser using DataLab