Intro

HashSet<T> stores unique values with O(1) average membership checks. Use it when uniqueness and lookup speed matter more than ordering. A concrete use case: deduplicating 500K event IDs in a message processing pipeline — HashSet<string>.Contains checks each ID in sub-microsecond time, while List<string>.Contains would scan up to 500K entries per check, turning a 200 ms job into a 40-minute one.

Deeper Explanation

HashSet<T> is hash-based: GetHashCode determines the bucket; Equals resolves collisions within a bucket. Core operations (Add, Contains, Remove) are O(1) average.

HashSet<T> also exposes the full set algebra in O(n) time:

• UnionWith — adds all elements from another collection (set A ∪ B).
• IntersectWith — keeps only elements present in both (A ∩ B).
• ExceptWith — removes elements found in another collection (A − B).
• SymmetricExceptWith — keeps elements present in exactly one set (A △ B).

These are in-place mutations. IsSubsetOf, IsSupersetOf, and SetEquals test structural relationships without modifying the set. A common production pattern: accumulate processed IDs in a HashSet<int>, then call ExceptWith against the full batch to find unprocessed items in O(n) instead of O(n²).

Structure

graph TD
    H1[value dotnet hash] --> B0[bucket zero]
    H2[value csharp hash] --> B1[bucket one]
    B0 --> V1[dotnet]
    B1 --> V2[csharp]

Example

var tags = new HashSet<string>(StringComparer.OrdinalIgnoreCase)
{
    "dotnet",
    "csharp"
};

var added = tags.Add("DOTNET"); // false, already exists by comparer

Pitfalls

• Broken GetHashCode/Equals contract — overriding Equals without matching GetHashCode causes Add to accept duplicates and Contains to miss existing entries. Two objects that are Equals must produce the same GetHashCode.
• Mutable fields in hash computation — mutating a field that participates in GetHashCode after insertion makes the entry unreachable in its bucket. Contains returns false even though the entry exists. Use immutable types or never mutate hash-participating fields after adding to the set.
• Assuming stable enumeration order — HashSet<T> enumeration order depends on internal bucket layout and can change after Add/Remove operations or across .NET versions. If you serialize a HashSet and compare output strings, tests will be flaky. Use SortedSet<T> or .OrderBy() when order matters.

Tradeoffs

• Use HashSet<T> for fast uniqueness checks.
• Use SortedSet<T> if you need sorted uniqueness and accept O(log n) operations.

Questions

Hash-Based Collections Comparison

Decision rule: use HashSet<T> when you only need to track membership or perform set operations (union, intersect, except). Use Dictionary when you need to associate a value with each key.

Links