Add Postman Sort (LSD Radix Sort) in R

sorting_algorithms/postman_sort.r

@@ -0,0 +1,59 @@
+# Postman sort (LSD Radix sort) in R:
+
+# This implements decimal LSD radix sort using a stable counting sort per digit.
+# It mirrors the provided Java version: repeatedly apply counting sort by digit
+# for exp = 1, 10, 100, ... up to the max number of digits.
+
+postman.sort <- function(elements.vec) {
+  if (!is.numeric(elements.vec)) stop("postman.sort: elements.vec must be numeric")
+  n <- length(elements.vec)
+  if (n <= 1) return(elements.vec)
+
+  # Only supports non-negative integers in standard form (typical radix LSD assumption)
+  if (any(elements.vec < 0)) stop("postman.sort: only non-negative values supported")
+
+  # For safety, coerce to integer-like values
+  if (any(elements.vec != floor(elements.vec))) stop("postman.sort: values must be integers")
+
+  get.max <- function(a) max(a)
+
+  counting.sort.by.digit <- function(a, exp) {
+    n <- length(a)
+    output <- numeric(n)
+    count <- integer(10) # digits 0..9
+
+    # Count occurrences
+    for (i in seq_len(n)) {
+      digit <- (a[i] %/% exp) %% 10
+      count[digit + 1L] <- count[digit + 1L] + 1L
+    }
+
+    # Prefix sums for positions (1-based indexing in R)
+    for (i in 2:10) {
+      count[i] <- count[i] + count[i - 1L]
+    }
+
+    # Build output stable, iterate from end
+    for (i in n:1) {
+      digit <- (a[i] %/% exp) %% 10
+      idx <- count[digit + 1L]
+      output[idx] <- a[i]
+      count[digit + 1L] <- idx - 1L
+    }
+
+    return(output)
+  }
+
+  max.val <- get.max(elements.vec)
+  exp <- 1
+  while ((max.val %/% exp) > 0) {
+    elements.vec <- counting.sort.by.digit(elements.vec, exp)
+    exp <- exp * 10
+  }
+
+  return(elements.vec)
+}
+
+# Example:
+# postman.sort(c(170, 45, 75, 90, 802, 24, 2, 66))
+# [1] 2 24 45 66 75 90 170 802