CareerCup

Hashmaps are better for average case search, generally O(c), where c is the number of elements in the longest linked list associated with an entry in a hashmap. (This is when we are making use of Chaining to avoid hash collision) We make sure that this happens with the help of load factor.

In the worst case scenario, c = n, i.e. there is only one chain and array and hashmap are practically equivalent.

If index is in the form of string then hashmap is better than array . hashmap perform very fast for large number , if it is small array and hashmap are same ... the time to calculate hash function has to be consider ..

worst case searching for Hash map is O(logn) because their internal implementation uses the B+/AVL Trees( It depends on compiler ). Other hand of arrays takes O(n)

Its all about the expected(realistic) performance. Same theory applied to bubble sort and qsort, both sorting are O(n^2) running time, but qsort is used everywhere since it's expected running time is o(N*lgN) unless the input data is sorted or almost sorted.

Array is a hashmap with 1-1 mapping, to save space we use hashmap
so if the data is limited and space is not a constrant then Array is better else
Hash map is the choice