18. Group Anagrams

mediumAsked at Coupang

Group strings that are anagrams of each other, mirroring how Coupang's Korean e-commerce search ranking clusters tokens that normalize to the same canonical form for retrieval.

By Alex Chen, Founder, InterviewChamp.AI · Last verified 2026-05-20

Problem

Given an array of strings strs, group the anagrams together. Return groups in any order.

Constraints

1 <= strs.length <= 10^4
0 <= strs[i].length <= 100
strs[i] consists of lowercase English letters

Examples

Example 1

Input

strs=["eat","tea","tan","ate","nat","bat"]

Output

[["eat","tea","ate"],["tan","nat"],["bat"]]

Example 2

Input

strs=[""]

Output

[[""]]

Approaches

1. Sorted-key bucketing

Sort each string's chars to form a canonical key; group by that key.

Time: O(n * k log k)
Space: O(n*k)

const m = new Map();
for (const s of strs) {
  const k = s.split('').sort().join('');
  if (!m.has(k)) m.set(k, []);
  m.get(k).push(s);
}
return [...m.values()];

Tradeoff:

2. Frequency-tuple bucketing

Build a 26-length count vector per string as the key; avoids sorting.

Time: O(n*k)
Space: O(n*k)

function groupAnagrams(strs) {
  const m = new Map();
  for (const s of strs) {
    const count = new Array(26).fill(0);
    for (const c of s) count[c.charCodeAt(0) - 97]++;
    const k = count.join(',');
    if (!m.has(k)) m.set(k, []);
    m.get(k).push(s);
  }
  return [...m.values()];
}

Tradeoff:

Coupang-specific tips

Coupang's Korean e-commerce search ranking clusters tokens normalized into a canonical form for retrieval; frequency-tuple bucketing scales better than sort-key under their throughput SLO.

Solve it now

Free. No sign-up. Python and JavaScript run instantly in your browser.

Output

Press Run or Cmd+Enter to execute

Practice these live with InterviewChamp.AI

Drill Group Anagrams and other Coupang interview questions under real-loop conditions with instant feedback on your reasoning, complexity claims, and code.

Practice these live with InterviewChamp.AI →