Why is a hash map preferred over an array for the "two sum" problem when aiming for O(n) time?

A A hash map allows O(1) average lookup of a complement value while scanning once, avoiding the O(n²) nested-loop comparison B A hash map automatically sorts the input, enabling binary search for the complement C A hash map uses less memory than an array of the same size D A hash map guarantees the result is found in the first half of the array