Design a Data Structure for Stock Prices (Heaps)

To solve this efficiently, I would design a system combining two priority queues with a hash map. First, I'd maintain a hash map where keys are stock symbols and values are their most recent prices. This ensures O(1) access to the current price. For the minimum and maximum queries, I would use a min-heap and a max-heap respectively, storing tuples of (price, timestamp, symbol). When recording a new price, I update the hash map and push the new entry into both heaps. Crucially, I cannot remove old entries from the heaps instantly without violating O(log n) constraints. Instead, I implement lazy deletion. When retrieving the maximum, I peek at the max-heap root. If the price there doesn't match the current value in my hash map, it's stale data, so I pop it and repeat until I find a valid entry. The same logic applies to the minimum. While individual operations might occasionally take longer due to cleanup, the amortized time complexity remains efficient. This approach mirrors Netflix's need for scalable real-time data processing, ensuring we get instant price checks while keeping order statistics robust without excessive memory churn from constant deletions.