Unordered associative containers (C++)

In the programming language C++, unordered associative containers are a group of class templates in the C++ Standard Library that implement hash table variants. Being templates, they can be used to store arbitrary elements, such as integers or custom classes. The following containers are defined in the current revision of the C++ standard: unordered_set, unordered_map, unordered_multiset, unordered_multimap. Each of these containers differ only on constraints placed on their elements.

C++26 introduces a hive data structure. It is essentially an object pool.^[1]

The unordered associative containers are similar to the associative containers in the C++ Standard Library but have different constraints. As their name implies, the elements in the unordered associative containers are not ordered. This is due to the use of hashing to store objects. The containers can still be iterated through like a regular associative container.

std::unordered_map and std::unordered_set are essentially (respectively) equivalent to java.util.HashMap and java.util.HashSet from Java, or std::collections::HashMap and std::collections::HashSet from Rust.

The first widely used implementation of hash tables in the C++ language was hash_map, hash_set, hash_multimap, hash_multiset class templates of the Silicon Graphics (SGI) Standard Template Library (STL).^[2] Due to their usefulness, they were later included in several other implementations of the C++ Standard Library (e.g., the GNU Compiler Collection's (GCC) libstdc++^[3] and the Visual C++ (MSVC) standard library).

The hash_* class templates were proposed into C++ Technical Report 1 (C++ TR1) and were accepted under names unordered_*.^[4] Later, they were incorporated into the C++11 revision of the C++ standard.^[5] An implementation is also available in the Boost C++ Libraries as <boost/unordered_map.hpp>.^[6]

The containers are defined in headers named after the names of the containers, e.g., unordered_set is defined in header <unordered_set>. All containers satisfy the requirements of the Container concept, which means they have begin(), end(), size(), max_size(), empty(), and swap() methods.

import std;

using std::string;
using std::unordered_map;

const unordered_map<string, int> MONTHS {
    {"January", 31},
    {"February", 28},
    {"March", 31},
    {"April", 30},
    {"May", 31},
    {"June", 30},
    {"July", 31},
    {"August", 31},
    {"September", 30},
    {"October", 31},
    {"November", 30},
    {"December", 31}
};
 
int main(int argc, char* argv[]) {
    std::println("September -> {}", MONTHS["September"]);
    std::println("April -> {}", MONTHS["April"]);
    std::println("December -> {}", MONTHS["December"]);
    std::println("February -> {}", MONTHS["February"]);
    return 0;
}

To use custom objects in std::unordered_map, a custom hasher must be defined. This function takes a const reference to the custom type and returns a size_t.

import std;

using std::hash;
 
struct Vector3 {
    int i;
    int j;
    int k;
};

struct HashVector3 {
    size_t operator()(const Vector3& x) const {
        return hash<int>()(x.i) ^ hash<int>()(x.j) ^ hash<int>()(x.k);
    }
};

The user defined function can be used as is in std::unordered_map, by passing it as a template parameter

unordered_map<Vector3, int, HashVector3> myPointToIntMap;

Or can be set as the default hash function by specializing std::hash:

namespace std {
    template <>
    class hash<Vector3> {
    public:
        size_t operator()(const Vector3& x) const {
            return hash<int>()(x.i) ^ hash<int>()(x.j) ^ hash<int>()(x.k);
        }
    };
}

//...
unordered_map<Vector3, int> myPointToIntMap;