Arrays.binarySearch vs Manual Binary Search Java | Fresher Interview

Arrays dot binarySearch is Java's built-in binary search for a sorted array. If the key is present it returns the index. The part freshers trip on is what it returns when the key is absent. It does not return minus one. It returns minus the insertion point minus one, where the insertion point is where the value would go to keep the array sorted. So if a missing value would belong at index two, it returns minus three. You decode the insertion point as minus the quantity return value plus one. That encoding is actually useful, because a single call tells you both that the value is missing and where it should be inserted. A manual binary search is the loop you write by hand, and I reach for it whenever I need something the library can't do. The classic cases are finding the first or last occurrence of a value that appears multiple times, searching in a rotated sorted array, or doing binary search on an answer space, like finding a minimum capacity, where there's no array to pass in at all. So my rule is simple. For a straightforward existence or index check on already-sorted data, I use the library method because it's one line and thoroughly tested. The moment the problem is a variant, I write the loop myself so I control exactly how the boundaries move. And in both cases the array has to be sorted first, and both run in O of log n time.