Data structures: which one when?

By codementor October 6, 2026

A one-page cheat sheet: what each basic data structure is good at, what it costs, and the C# type to use.

Data structures: which one when?

Pick a data structure by asking: what will I do most often? Read by position? Look up by key? Add and remove at the ends? Keep things sorted?

Need	Use	C# type	Main cost
Read by position, loop over all	Array / list	`int[]`, `List<T>`	index O(1), middle insert O(n)
Look up by key, count, remove duplicates	Hash table	`Dictionary<K,V>`, `HashSet<T>`	O(1) on average
Last in, first out (undo, brackets, DFS)	Stack	`Stack<T>`	push amortized O(1), pop O(1)
First in, first out (BFS, jobs)	Queue	`Queue<T>`	enqueue amortized O(1), dequeue O(1)
Always take the smallest (top k, Dijkstra)	Heap	`PriorityQueue<T,P>`	O(log n)
Keep sorted while changing, ranges	Balanced tree	`SortedSet<T>`, `SortedDictionary<K,V>`	O(log n)
Many inserts/removes at known nodes	Linked list	`LinkedList<T>`	O(1) at a node, O(n) to find it
Grid data	Matrix	`int[,]`	O(rows × cols) to visit all
Nested data	Tree	your own node class	depends on shape

How to answer in an interview

Say what operation the problem needs most.
Pick the structure that makes that operation cheap.
Say the cost out loud, and the trade-off: a hash table is fast but unordered; a sorted set is ordered but O(log n).

A quick example

“Find if any two numbers sum to 10” needs fast “have I seen x?” checks, so a HashSet<int> turns it into one pass:

int[] nums = { 3, 9, 7, 1 };
var seen = new HashSet<int>();
bool found = false;
foreach (int x in nums)
{
    found |= seen.Contains(10 - x);
    seen.Add(x);
}
Console.WriteLine(found);

It prints True (3 + 7 = 10).

#tip · TIP-032

#tip