C++ Hashmap (2.3.0)

Table of Contents

• About
• Instructions
• Dependencies
• References

About

This is a hashmap written in C++.

It has a similar API to C#'s Dictionary and self-initializes like an std::vector. Currently, it uses sequential chaining for collision resolution. More collision resolution techniques may be added in the future.

This structure provides robust exception safety, and is suitable for use in a concurrent environment. Furthermore, it supports move semantics and initialiser lists.

Explicit finalization of the hashmap is not necessary. However, if you are storing manually-managed memory, then remember to free any elements before removal.

If you find a bug or have a feature-request, please raise an issue.

Like hashsets? Check out my other project: cpp-hashset!

Instructions

The implementation is header-only and written in templated C++17. You should not need to make any adjustments to your project settings or compiler flags.

Simply include it in your project and you are ready to start!

Example:

#include <string>

#include "Hashmap.hpp"

LouiEriksson::Hashmap<std::string, float> hashmap {
    { "key1", 1.0f },
    { "key2", 2.0f },
    { "key3", 3.0f },
}

int main() {

    if (const auto item = hashmap.Get("key3")) {
        std::cout << "Value: " << item.value() << '\n';
    }
    else {
        std::cout << "Key not in Hashmap!\n";
    }

    return 0;
}

Dependencies

The hashmap was written in C++17 and utilises the following standard headers:

<algorithm>

<cstddef>

<functional>

<initializer_list>

<iostream>

<mutex>

<optional>

<stdexcept>

<vector>

Why not use <unordered_map>?

I find unordered_map to be way too verbose for most situations.

In this implementation, key existence and value retrieval are merged into a single conditional expression. This allows for simpler, cleaner code that affords better exception handling.

Note

Please note that while the hashmap is capable of being used in a concurrent environment, it does not provide a mechanism for synchronising changes to the hashmap which are made in between operations.

Therefore, if you need to perform a synchronous series of operations on the Hashmap while it is being used in a concurrent context, you should lock access to the hashmap to one thread while doing so, otherwise you may encounter race conditions.

References

• Wang, Q. (Harry) (2020). Implementing Your Own HashMap (Explanation + Code). YouTube. Available at: https://www.youtube.com/watch?v=_Q-eNqTOxlE [Accessed 2021].