First non-repeating character in a stream of data

For each incoming character, increment its frequency in a map and append it to a queue. Then repeatedly remove the queue front while that character's frequency is greater than one. The remaining front, if any, is the first character that has appeared exactly once because the queue preserves arrival order and stale repeated candidates are removed only when they reach the front. This lazy cleanup is simpler than searching all seen characters after every input. Clarify whether input is bytes, Unicode code points, or grapheme clusters; a fixed array is safe only for a defined alphabet. Test a stream beginning with repeats, a later unique character, all repeats, and spaces or punctuation if accepted. Each character enters once and leaves once, so total work over n arrivals is O(n), amortized O(1) per arrival, with O(k) space for k distinct tracked characters.