Skip to content

Incorrect caching of recursive functions #158

Description

@venpopov

If you memoise a recursive function, like the factorial function below, and then call the function in a loop or sapply, you get duplicate cache keys. For example, here we get 19 keys in the cache instead of 10:

library(memoise)

cache_size <- function(fun) {
  env <- environment(fun)
  cache <- env[["_cache"]]
  cache$size()
}

factorial <- function(n) {
  if (n == 1) return(1)
  n * factorial(n-1)
}

factorial <- memoise(factorial)

cache_size(factorial)
#> [1] 0

for (i in 1:10) {
  a <- factorial(i)
}

cache_size(factorial)
#> [1] 19

Created on 2026-02-13 with reprex v2.1.1

Created on 2026-02-13 with reprex v2.1.1

If instead you call the function manually with all the arguments, you get the correct number:

library(memoise)

cache_size <- function(fun) {
  env <- environment(fun)
  cache <- env[["_cache"]]
  cache$size()
}

factorial <- function(n) {
  if (n == 1) return(1)
  n * factorial(n-1)
}

factorial <- memoise(factorial)

cache_size(factorial)
#> [1] 0

a <- factorial(1)
a <- factorial(2)
a <- factorial(3)
a <- factorial(4)
a <- factorial(5)
a <- factorial(6)
a <- factorial(7)
a <- factorial(8)
a <- factorial(9)
a <- factorial(10)

cache_size(factorial)
#> [1] 10

Created on 2026-02-13 with reprex v2.1.1

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions