std::pair

std::pair stores exactly two values, first and second, as one object.

# Semantics

• pair<T1, T2> is the two-element building block in the tuple family: fixed size, heterogeneous, and value-oriented.
• The two elements are named members, not positional access only, which makes pair suitable when the values naturally form a small named association.
• Unlike std::tuple, pair is intentionally specialized for arity two and integrates directly with tuple-like APIs such as get, tuple_size, and tuple_element.
• pair is commonly used for return values, map-like key/value objects, and algorithms that naturally produce two related results.

Use pair when the two-field grouping is obvious and stable. If the object needs domain-specific meaning, a named struct is often clearer.

# Declarations

template< class T1, class T2 >
struct pair;

# Member objects

# Example

#include <iostream>
#include <string>
#include <utility>
 
std::pair<std::string, int> describe()
{
    return {"apples", 3};
}
 
int main()
{
    auto [name, count] = describe();
    std::cout << name << ": " << count << '\n';
}

This is the common pattern: bundle two related results together, then unpack them with structured bindings or get.

# Operational notes

• Construction and assignment are covered by pair::pair and pair::operator=.
• swap exchanges the two element values with another pair of the same type.
• make_pair is useful when you want type deduction without spelling pair<T1, T2> explicitly.
• comparison operators compare pairs lexicographically: first is compared before second.
• get lets pair participate in tuple-like generic code.

# Reference map

# Notes

If neither T1 nor T2 is a possibly cv-qualified class type with a non-trivial destructor, or an array of such a type, the destructor of pair is trivial.

# Defect reports

# See also

• std::tuple
• std::tie