15. LFU Cache
hardAsked at RedisImplement a Least-Frequently-Used cache with O(1) get and put; Redis loves it because LFU eviction (allkeys-lfu) is one of its native maxmemory policies.
By Alex Chen, Founder, InterviewChamp.AI · Last verified
Problem
Implement LFUCache with capacity C. Both get and put must run in O(1). On capacity overflow evict the least-frequently-used key; break ties by LRU among that frequency tier.
Constraints
0 <= capacity <= 10^4Up to 2 * 10^5 calls
Examples
Example 1
capacity=2; put(1,1); put(2,2); get(1)=1; put(3,3) evicts 2; get(2)=-1see explanationApproaches
1. Sort by frequency on eviction
Maintain a Map of {key: [value, freq]} and scan on each put.
- Time
- O(n) per put
- Space
- O(n)
// Scan map for (minFreq, oldestAtMinFreq) on every eviction.Tradeoff:
2. Three maps: key->node, freq->list, plus minFreq
Store each key in a doubly linked list bucketed by frequency. On get bump the node to the next-higher freq list. Maintain minFreq for O(1) eviction. Mirrors Redis's LFU counter with logarithmic-decay sampling.
- Time
- O(1) per op
- Space
- O(capacity)
class LFUCache {
constructor(cap) {
this.cap = cap;
this.size = 0;
this.minFreq = 0;
this.keyNode = new Map();
this.freqList = new Map(); // freq -> LinkedList of nodes
}
get(k) { /* if missing return -1; else bump freq and return value */ }
put(k, v) { /* if exists update value + bump; else evict if full, insert with freq=1, minFreq=1 */ }
}Tradeoff:
Redis-specific tips
Redis interviewers grade for the three-map design AND for naming the LFU counter decay trick — real Redis uses an 8-bit counter that probabilistically increments and decays over time to avoid frequency-stuck keys.
Solve it now
Free. No sign-up. Python and JavaScript run instantly in your browser.
Practice these live with InterviewChamp.AI
Drill LFU Cache and other Redis interview questions under real-loop conditions with instant feedback on your reasoning, complexity claims, and code.
Practice these live with InterviewChamp.AI →