Difference between revisions of "cpp/utility/hash"

Revision as of 10:08, 11 July 2020

Each specialization of this template is either enabled ("untainted") or disabled ("poisoned"). For every type Key for which neither the library nor the user provides an enabled specialization std::hash<Key>, that specialization exists and is disabled. Disabled specializations do not satisfy Hash, do not satisfy FunctionObject, and following values are all false:

• std::is_default_constructible<std::hash<Key>>::value
• std::is_copy_constructible<std::hash<Key>>::value
• std::is_move_constructible<std::hash<Key>>::value
• std::is_copy_assignable<std::hash<Key>>::value
• std::is_move_assignable<std::hash<Key>>::value

In other words, they exist, but cannot be used.

The enabled specializations of the hash template defines a function object that implements a hash function. Instances of this function object satisfy Hash. In particular, they define an operator() const that:

1. Accepts a single parameter of type Key.
2. Returns a value of type std::size_t that represents the hash value of the parameter.
3. Does not throw exceptions when called.
4. For two parameters k1 and k2 that are equal, std::hash<Key>()(k1) == std::hash<Key>()(k2).
5. For two different parameters k1 and k2 that are not equal, the probability that std::hash<Key>()(k1) == std::hash<Key>()(k2) should be very small, approaching 1.0/std::numeric_limits<std::size_t>::max().

All explicit and partial specializations of hash provided by the standard library are DefaultConstructible, CopyAssignable, Swappable and Destructible. User-provided specializations of hash also must meet those requirements.

The unordered associative containers std::unordered_set, std::unordered_multiset, std::unordered_map, std::unordered_multimap use specializations of the template std::hash as the default hash function.

Notes

The actual hash functions are implementation-dependent and are not required to fulfill any other quality criteria except those specified above. Notably, some implementations use trivial (identity) hash functions which map an integer to itself. In other words, these hash functions are designed to work with unordered associative containers, but not as cryptographic hashes, for example.

Hash functions are only required to produce the same result for the same input within a single execution of a program; this allows salted hashes that prevent collision denial-of-service attacks.

There is no specialization for C strings. std::hash<const char*> produces a hash of the value of the pointer (the memory address), it does not examine the contents of any character array.

Member functions

Standard specializations for basic types

In addition to the above, the standard library provides specializations for all (scoped and unscoped) enumeration types. These may be (but are not required to be) implemented as std::hash<std::underlying_type<Enum>::type>)

The standard library provides enabled specializations of std::hash for std::nullptr_t and all cv-unqualified arithmetic types (including any extended integer types), all enumeration types, and all pointer types.

Each standard library header that declares the template std::hash provides all enabled specializations described above. These headers include <string>, <system_error>, <bitset>, <memory>, <typeindex>, <vector>, <thread>, <optional>, <variant>, <string_view>, <coroutine>(since C++20).

Standard specializations for library types

Note: additional specializations for std::pair and the standard container types, as well as utility functions to compose hashes are available in boost.hash

Example

#include <iostream>
#include <iomanip>
#include <functional>
#include <string>
#include <unordered_set>
 
struct S {
    std::string first_name;
    std::string last_name;
};
bool operator==(const S& lhs, const S& rhs) {
    return lhs.first_name == rhs.first_name && lhs.last_name == rhs.last_name;
}
 
// custom hash can be a standalone function object:
struct MyHash
{
    std::size_t operator()(S const& s) const noexcept
    {
        std::size_t h1 = std::hash<std::string>{}(s.first_name);
        std::size_t h2 = std::hash<std::string>{}(s.last_name);
        return h1 ^ (h2 << 1); // or use boost::hash_combine (see Discussion)
    }
};
 
// custom specialization of std::hash can be injected in namespace std
namespace std
{
    template<> struct hash<S>
    {
        std::size_t operator()(S const& s) const noexcept
        {
            std::size_t h1 = std::hash<std::string>{}(s.first_name);
            std::size_t h2 = std::hash<std::string>{}(s.last_name);
            return h1 ^ (h2 << 1); // or use boost::hash_combine
        }
    };
}
 
int main()
{
 
    std::string str = "Meet the new boss...";
    std::size_t str_hash = std::hash<std::string>{}(str);
    std::cout << "hash(" << std::quoted(str) << ") = " << str_hash << '\n';
 
    S obj = { "Hubert", "Farnsworth"};
    // using the standalone function object
    std::cout << "hash(" << std::quoted(obj.first_name) << ',' 
               << std::quoted(obj.last_name) << ") = "
               << MyHash{}(obj) << " (using MyHash)\n                           or "
               << std::hash<S>{}(obj) << " (using injected std::hash<S> specialization)\n";
 
    // custom hash makes it possible to use custom types in unordered containers
    // The example will use the injected std::hash<S> specialization above,
    // to use MyHash instead, pass it as a second template argument
    std::unordered_set<S> names = {obj, {"Bender", "Rodriguez"}, {"Turanga", "Leela"} };
    for(auto& s: names)
        std::cout << std::quoted(s.first_name) << ' ' << std::quoted(s.last_name) << '\n';
}

hash("Meet the new boss...") = 1861821886482076440
hash("Hubert","Farnsworth") = 17622465712001802105 (using MyHash)
                           or 17622465712001802105 (using injected std::hash<S> specialization) 
"Turanga" "Leela"
"Bender" "Rodriguez"
"Hubert" "Farnsworth"

Defect reports

The following behavior-changing defect reports were applied retroactively to previously published C++ standards.