Read this blog: https://codeforces.net/blog/entry/62393

There is performance difference between std::unordered_map and std::map in terms of finding elements and to understand this you need to know the internal implementation of these data Structures.

1. std::unordered_map:

• std::unordered_map is implemented as a hash table, where elements are stored in buckets based on their hash values. When you perform a find operation in an unordered_map, it has an average time complexity of O(1), assuming a good hash function and a low collision rate.
• However, in some worst-case scenarios, particularly when there are many hash collisions, the time complexity can degrade to O(n), where n is the number of elements in the map. This happens because all elements in the same bucket need to be examined linearly to resolve the collision.

1. std::map (ordered_map):

• std::map (or std::ordered_map) is implemented as a balanced binary search tree (usually a Red-Black Tree). This data structure maintains the elements in sorted order based on their keys.
• The find operation in a std::map has a time complexity of O(log n), where n is the number of elements in the map. This is because it performs a binary search on the tree structure to locate the element.

So, why might you observe a TLE (Time Limit Exceeded) when using std::unordered_map::find but not when using std::map::find?

The most likely explanation is that your std::unordered_map is experiencing worst-case behavior due to hash collisions. In cases where there are many hash collisions (e.g., if your hash function is not well-distributed or if you have a lot of elements with the same hash), the std::unordered_map can degrade in performance and, in the worst case, become slower than a std::map for finding elements. This can lead to TLE errors.