What are the essential data structures?
A handful of structures cover most backend work: arrays and lists for ordered data, maps and objects for key lookups, sets for uniqueness, and queues and stacks for ordering work. Picking the right one turns slow, awkward code into fast, obvious code.
Why it matters
The wrong structure is a silent performance bug. Scanning an array to check membership is fine for ten items and a disaster for a million — a set does it instantly. Backend code runs on real volumes, so the structure you reach for decides whether a request takes a millisecond or a second. Interviews probe this directly.
What to learn
- Big-O in plain terms: how cost grows with input size
- Arrays vs linked lists, and the cost of insert vs index
- Maps and objects for O(1) lookups by key
- Sets for membership and de-duplication
- Stacks and queues for ordered processing
- When a tree or heap is worth it
- Choosing a structure from the access pattern, not habit
Common pitfall
Reaching for an array and .includes() for membership checks inside a loop.
That is O(n) per check and O(n squared) overall. If you are repeatedly asking
"have I seen this?", use a Set; if you are looking things up by key, use a
Map. The fix is usually one line and a big speedup.
Resources
Primary (free):
- MDN — Map · docs
- MDN — Set · docs
- JavaScript.info — Data structures · docs
Practice
Take a list of 100k duplicate records and de-duplicate them two ways: once with
an array and .includes(), once with a Set. Time both. Then build a lookup
that maps user ids to records with a Map and fetch by id. Done when you can
explain the timing difference in Big-O terms.
Outcomes
- State the lookup and insert cost of arrays, maps, and sets.
- Pick the right structure from the access pattern.
- Replace an O(n squared) membership check with a set.
- Estimate how code will behave as data volume grows.