Playing around with reduce, Reduce, and foldr1

library(tidyverse)

set.seed(20250104)
A <- data.frame(id = 1:3, x = rnorm(3))
B <- data.frame(id = c(2, 3), y = rnorm(2))
C <- data.frame(id = c(1, 3), z = rnorm(2))

A

B

C

Alexander Krannich’s example:

list(A, B, C) |>
  reduce(full_join, by = "id")

Using only {base} (mild edit of David Souza’s reply):

list(A, B, C) |>
  Reduce(\(x, y) merge(x, y, by = "id", all = TRUE), x = _)

Getting silly – this all reminded me of Haskell’s foldr1, implemented (without the Currying) as:

foldr1 <- \(f, xs) {
  if (length(xs) == 1)
    xs[[1]]
  else
    f(xs[[1]], foldr1(f, tail(xs, length(xs) - 1)))
}

foldr1(\(x, y) merge(x, y, by = "id", all = TRUE), list(A, B, C))

Here is an easier example:

foldr1(`+`, 1:10)

[1] 55

Which is the same as:

sum(1:10)

[1] 55