HashMap vs ConcurrentHashMap — what’s the practical difference?

Deep dive

`HashMap` has no concurrency guarantees. Concurrent writes can lead to lost updates and even internal corruption.

`ConcurrentHashMap` is designed for high concurrency:

• Updates use fine-grained synchronization/CAS on bins.
• Iterators are weakly consistent (they don’t throw ConcurrentModificationException and may not see all updates).
• It disallows null keys/values (null would be ambiguous for “missing”).

Practical guidance

• Many threads, reads + writes: use `ConcurrentHashMap`.
• You still need atomic operations for compound actions (check-then-act): prefer `computeIfAbsent`, `merge`, `compute`.

Common pitfalls

• Thinking it makes *your whole algorithm* thread-safe.
• Using `get()` then `put()` as if it were atomic.
• Putting nulls (not allowed).