std::unique(3)

💅 std::unique — the C++ Standard Libary one, bestie

Okay so this is std::unique(3) from the C++ Standard Libary, and honestly? She's iconic but deeply misunderstood. Let me put you on. ✨

---

🌸 NAME 🌸

std::unique — yes bestie, the name is literally just std::unique, that's the whole thing, no cap.

---

✨ Synopsis ✨

Defined in header <algorithm>:

template< class ForwardIt >                              (1)
(constexpr since C++20)
ForwardIt unique( ForwardIt first, ForwardIt last );

template< class ExecutionPolicy, class ForwardIt >
ForwardIt unique( ExecutionPolicy&& policy,              (2) (since C++17)
                  ForwardIt first, ForwardIt last );

template< class ForwardIt, class BinaryPred >
ForwardIt unique( ForwardIt first, ForwardIt last,       (3)
(constexpr since C++20)
                  BinaryPred p );

template< class ExecutionPolicy,
          class ForwardIt, class BinaryPred >
ForwardIt unique( ExecutionPolicy&& policy,              (4) (since C++17)
                  ForwardIt first, ForwardIt last,
                  BinaryPred p );

So what does she do? She removes all except the first element from every consecutive group of equivalent elements from the range [first, last), and then hands you back a past-the-end iterator for the new end of the range. Slay. 🫶

• Overload 1) compares elements using operator==. And listen — if operator== does not establish an equivalence relation, the behavior is undefined. Not joking, not cute, undefined. 💀
• Overload 3) compares using the binary predicate p you give her. Same rule though: if p doesn't establish an equivalence relation, behavior is undefined.
• Overloads 2) and 4) are the same as (1) and (3) but executed according to policy. These overloads participate in overload resolution only if std::is_execution_policy_v<std::decay_t<ExecutionPolicy>> is true (until C++20), or std::is_execution_policy_v<std::remove_cvref_t<ExecutionPolicy>> is true (since C++20).

---

🦋 Explanation 🦋

The removing is done by shifting the elements in the range so that the elements that are not to be removed appear at the beginning of the range. Think of it like shoving the keepers to the front of the line. 🎀

• Shifting is done by copy assignment (until C++11) or move assignment (since C++11).
• The removing operation is stable: the relative order of the elements not to be removed stays the same. Periodt.
• The underlying sequence of [first, last) is not shortened by the removing operation. Given result as the returned iterator:
  • All iterators in [result, last) are still dereferenceable.
  • Each element of [result, last) has a valid but unspecified state, because move assignment can eliminate elements by moving from elements that were originally in that range (since C++11).

---

💜 Parameters 💜

• first, last — the range of elements to process.
• policy — the execution policy to use. See execution policy for details.
• p — binary predicate which returns true if the elements should be treated as equal.

The signature of the predicate function should be equivalent to the following:

bool pred(const Type1 &a, const Type2 &b);

And about p, ngl there are rules:

• While the signature does not need to have const &, the function must not modify the objects passed to it.
• It must be able to accept all values of type (possibly const) Type1 and Type2 regardless of value category — thus Type1 & is not allowed, nor is Type1 unless for Type1 a move is equivalent to a copy (since C++11).
• The types Type1 and Type2 must be such that an object of type ForwardIt can be dereferenced and then implicitly converted to both of them.

---

🧚 Type requirements 🧚

• ForwardIt must meet the requirements of LegacyForwardIterator.
• The type of dereferenced ForwardIt must meet the requirements of MoveAssignable.

---

👑 Return value 👑

A ForwardIt to the new end of the range. That's it, that's the vibe check. ⭐

---

🔥 Complexity 🔥

Given N as std::distance(first, last):

• 1, 2) Exactly max(0,N-1) comparisons using operator==.
• 3, 4) Exactly max(0,N-1) applications of the predicate p.

---

🙈 Exceptions 🙈

The overloads with a template parameter named ExecutionPolicy report errors as follows:

• If execution of a function invoked as part of the algorithm throws an exception and ExecutionPolicy is one of the standard policies, std::terminate is called. For any other ExecutionPolicy, the behavior is implementation-defined. 😭
• If the algorithm fails to allocate memory, std::bad_alloc is thrown.

---

🌈 Possible implementation 🌈

See also the implementations in libstdc++, libc++, and MSVC STL.

unique (1):

template<class ForwardIt>
ForwardIt unique(ForwardIt first, ForwardIt last)
{
    if (first == last)
        return last;

    ForwardIt result = first;
    while (++first != last)
        if (!(*result == *first) && ++result != first)
            *result = std::move(*first);

    return ++result;
}

unique (3):

template<class ForwardIt, class BinaryPredicate>
ForwardIt unique(ForwardIt first, ForwardIt last, BinaryPredicate p)
{
    if (first == last)
        return last;

    ForwardIt result = first;
    while (++first != last)
        if (!p(*result, *first) && ++result != first)
            *result = std::move(*first);

    return ++result;
}

---

🍓 Notes 🍓

A call to unique is typically followed by a call to a container's erase member function to actually remove elements from the container. Like, unique shifts stuff around but she does not shrink your container — you gotta do the erase yourself, bestie. 💅

---

💖 Example 💖

// Run this code

#include <algorithm>
#include <iostream>
#include <vector>

int main()
{
    // a vector containing several duplicate elements
    std::vector<int> v{1, 2, 1, 1, 3, 3, 3, 4, 5, 4};
    auto print = [&](int id)
    {
        std::cout << "@" << id << ": ";
        for (int i : v)
            std::cout << i << ' ';
        std::cout << '\n';
    };
    print(1);

    // remove consecutive (adjacent) duplicates
    auto last = std::unique(v.begin(), v.end());
    // v now holds {1 2 1 3 4 5 4 x x x}, where 'x' is indeterminate
    v.erase(last, v.end());
    print(2);

    // sort followed by unique, to remove all duplicates
    std::sort(v.begin(), v.end()); // {1 1 2 3 4 4 5}
    print(3);

    last = std::unique(v.begin(), v.end());
    // v now holds {1 2 3 4 5 x x}, where 'x' is indeterminate
    v.erase(last, v.end());
    print(4);
}

Output:

@1: 1 2 1 1 3 3 3 4 5 4
@2: 1 2 1 3 4 5 4
@3: 1 1 2 3 4 4 5
@4: 1 2 3 4 5

---

🎀 Defect reports 🎀

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

• LWG 202, applied to C++98: the behavior was unclear if the elements are compared using a non-equivalence relation → the behavior is undefined in this case.

---

🌸 See also 🌸

• adjacent_find — finds the first two adjacent items that are equal (or satisfy a given predicate) (function template).
• unique_copy — creates a copy of some range of elements that contains no consecutive duplicates (function template).
• remove, remove_if — removes elements satisfying specific criteria (function template).
• unique — removes consecutive duplicate elements (public member function of std::list<T,Allocator>).
• unique — removes consecutive duplicate elements (public member function of std::forward_list<T,Allocator>).
• ranges::unique (C++20) — removes consecutive duplicate elements in a range (niebloid).

---

Okay that's the whole page, mwah. 💜 This one's from cppreference.com, dated 2024.06.10 — go drink some water and then go call erase, bestie. 🦋