How do you choose the right collection? Compare Set/Map implementations and their ordering guarantees.

🎤 Say this first

Decide on three axes: ordering (none → Hash, insertion/access → LinkedHash, sorted → Tree), concurrency (single-thread → plain, concurrent → Concurrent/CopyOnWrite), and special keys (enums → EnumMap/EnumSet, priorities → PriorityQueue). Tree costs O(log n) but gives range queries and ordering; Hash* is O(1) but unordered. Choosing the tightest structure documents intent as much as it buys performance.

The full picture

Need	Use	Why
Fast lookup, no order	`HashMap` / `HashSet`	O(1) expected
Predictable iteration order	`LinkedHashMap` / `LinkedHashSet`	hash speed + insertion order; LRU via access-order
Sorted keys, range queries	`TreeMap` / `TreeSet`	red-black tree; `headMap`, `ceilingKey`, `subSet`
Enum keys	`EnumMap` / `EnumSet`	array/bitset backed — fastest and smallest
Queue/stack both ends	`ArrayDeque`	ring buffer; beats Stack & LinkedList
Priority scheduling	`PriorityQueue`	binary heap; note: iteration is NOT sorted
Read-mostly listeners	`CopyOnWriteArrayList`	snapshot iteration, zero read locking

Two habits mark a senior: (1) program to the interface — return List/Map/Collection from APIs, keep the implementation swappable; (2) immutability at boundaries — expose List.copyOf(...)/unmodifiable views so callers can't mutate your internals. Also know that TreeMap requires a total order consistent with equals, or Set semantics quietly break.

🎤 Say this first

The full picture

🔄 Likely follow-up questions