#####   ######  #####    ###    #   #  ###  #   #  ######
##  ##  ##      ##  ##  ## ##   #   #   #   #   #  ##
#####   ####    #####   #   #   #   #   #   #   #  ####
##  #   ##      ##      ## ##    # #    #    # #   ##
##   #  ######  ##       ###      #    ###    #    ######

$ curl repovive.com/roadmaps/graph-theory/lowest-common-ancestor-lca/implementation-counting-paths

░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░░██████████████████████████████████████████████████████████████████████████████████████████████████

#####   ######  #####    ###    #   #  ###  #   #  ######
##  ##  ##      ##  ##  ## ##   #   #   #   #   #  ##
#####   ####    #####   #   #   #   #   #   #   #  ####
##  #   ##      ##      ## ##    # #    #    # #   ##
##   #  ######  ##       ###      #    ###    #    ######

$ curl repovive.com/roadmaps/graph-theory/lowest-common-ancestor-lca/implementation-counting-paths

Repovive

Implementation - Counting Paths - Lowest Common Ancestor (LCA) | Graph Theory | Repovive

Repovive.

Graph Theory37 sections · 1633 units81h total

Open in Course

Implementation - Counting Paths

(Pseudocode)

cnt = array of size n, initialized to 0

For each path (u, v):
    lca_node = lca(u, v)
    cnt[u] += 1
    cnt[v] += 1
    cnt[lca_node] -= 2

function propagate(node):
    for each child c of node:
        propagate(c)
        cnt[node] += cnt[c]

propagate(root)
Output cnt[]

Time: $O(m \log n + n)$ with binary lifting for LCA, $O(n)$ for DFS propagation.

The key: marking is $O(1)$ per path, propagation is $O(n)$ total. Much better than walking each path individually, which would be $O(m \cdot n)$ .