Updated Jun 4, 2026

Ranges

Ranges (C++20) is a modern library that provides composable, lazy-evaluated operations on sequences. It replaces traditional iterator pairs with range objects and introduces views for efficient data transformation pipelines.

Why Ranges?

Traditional STL algorithms require iterator pairs, which is verbose and error-prone. Ranges simplify this:

#include <vector>
#include <algorithm>
#include <ranges>

std::vector<int> nums = {1, 2, 3, 4, 5};

// Old way: iterator pairs
std::sort(nums.begin(), nums.end());

// New way: ranges
std::ranges::sort(nums);

Benefits:

More concise syntax
Composable operations (pipelines)
Lazy evaluation (no intermediate containers)
Better compile-time error messages

Range Concepts

Ranges are defined by concepts that specify requirements:

#include <ranges>
#include <vector>
#include <list>

template<std::ranges::range R>
void process(R&& r) {
    // Works with any range
}

// Different range types
std::vector<int> vec;        // random_access_range
std::list<int> lst;          // bidirectional_range
std::istream_iterator<int> it; // input_range

Views

Views are lightweight, lazy-evaluated range adaptors. They don't own data and operations are performed on-demand.

Basic Views

#include <ranges>
#include <vector>
#include <iostream>

void basicViews() {
    std::vector<int> nums = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10};
    
    // Filter: select elements
    auto evens = nums | std::views::filter([](int n) { return n % 2 == 0; });
    
    // Transform: modify elements
    auto squared = nums | std::views::transform([](int n) { return n * n; });
    
    // Take: first N elements
    auto first3 = nums | std::views::take(3);
    
    // Drop: skip first N elements
    auto skip2 = nums | std::views::drop(2);
    
    // Views are lazy - no computation yet!
    for (int n : evens) {
        std::cout << n << ' ';  // 2 4 6 8 10
    }
}

info

Views are lazy - they only compute values when iterated. This means no intermediate allocations!

View Composition (Pipelines)

The power of ranges comes from composing multiple views:

#include <ranges>
#include <vector>

void pipelineExample() {
    std::vector<int> nums = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10};
    
    // Chain multiple operations
    auto result = nums
        | std::views::filter([](int n) { return n % 2 == 0; })  // Keep evens
        | std::views::transform([](int n) { return n * n; })     // Square them
        | std::views::take(3);                                   // First 3
    
    // Prints: 4 16 36
    for (int n : result) {
        std::cout << n << ' ';
    }
}

Common Views

View	Purpose	Example
`filter`	Select elements by predicate	Evens only
`transform`	Apply function to each element	Square values
`take`	First N elements	First 5 items
`take_while`	Take until condition fails	While < 100
`drop`	Skip first N elements	Skip header
`drop_while`	Skip until condition fails	Skip negatives
`reverse`	Reverse order	Backwards
`keys`	Extract keys from pairs	Map keys
`values`	Extract values from pairs	Map values

#include <ranges>
#include <map>

void commonViewsExample() {
    std::map<int, std::string> data = {{1, "one"}, {2, "two"}, {3, "three"}};
    
    // Extract keys
    auto keys = data | std::views::keys;
    
    // Extract values
    auto values = data | std::views::values;
    
    // Reverse iteration
    std::vector<int> nums = {1, 2, 3, 4, 5};
    auto reversed = nums | std::views::reverse;
}

Range Factories

Create ranges from scratch without underlying containers:

#include <ranges>
#include <iostream>

void rangeFactories() {
    // iota: sequence of incrementing values
    auto numbers = std::views::iota(1, 11);  // 1, 2, ..., 10
    
    // Infinite range (be careful!)
    auto infinite = std::views::iota(0);  // 0, 1, 2, 3, ...
    auto first10 = infinite | std::views::take(10);
    
    // single: single-element range
    auto one = std::views::single(42);
    
    // empty: empty range
    auto nothing = std::views::empty<int>;
    
    // repeat: same value N times (C++23)
    // auto fives = std::views::repeat(5, 10);  // 5, 5, 5, ... (10 times)
}

warning

std::views::iota() without a second argument creates an infinite range. Always use take() or similar to limit it!

Range Algorithms

C++20 adds range-based versions of all STL algorithms:

#include <ranges>
#include <vector>
#include <algorithm>

void rangeAlgorithms() {
    std::vector<int> nums = {3, 1, 4, 1, 5, 9, 2, 6};
    
    // Sort entire range
    std::ranges::sort(nums);
    
    // Find element
    auto it = std::ranges::find(nums, 5);
    
    // Count occurrences
    auto count = std::ranges::count(nums, 1);
    
    // Copy with projection
    std::vector<int> dest(nums.size());
    std::ranges::copy(nums, dest.begin());
    
    // Unique elements
    auto [first, last] = std::ranges::unique(nums);
    nums.erase(first, last);
}

Projections

Projections allow you to specify how to extract values for comparison:

#include <ranges>
#include <vector>
#include <algorithm>

struct Person {
    std::string name;
    int age;
};

void projectionExample() {
    std::vector<Person> people = {
        {"Alice", 30},
        {"Bob", 25},
        {"Charlie", 35}
    };
    
    // Sort by age using projection
    std::ranges::sort(people, {}, &Person::age);
    
    // Sort by name
    std::ranges::sort(people, {}, &Person::name);
    
    // Find person by age
    auto it = std::ranges::find(people, 25, &Person::age);
}

Projections

Projections eliminate the need for custom comparators in many cases!

Practical Examples

Example 1: Processing Text Lines

#include <ranges>
#include <string>
#include <vector>
#include <iostream>

void processTextLines() {
    std::vector<std::string> lines = {
        "  hello  ",
        "world",
        "",
        "  C++20  ",
        "ranges"
    };
    
    auto result = lines
        | std::views::filter([](const auto& s) { return !s.empty(); })
        | std::views::transform([](auto s) {
              // Trim spaces (simplified)
              return s;
          })
        | std::views::take(3);
    
    for (const auto& line : result) {
        std::cout << line << '\n';
    }
}

Example 2: Mathematical Operations

#include <ranges>
#include <vector>
#include <numeric>

void mathOperations() {
    std::vector<int> nums = {1, 2, 3, 4, 5};
    
    // Sum of squares of even numbers
    auto sum = std::ranges::fold_left(
        nums 
        | std::views::filter([](int n) { return n % 2 == 0; })
        | std::views::transform([](int n) { return n * n; }),
        0,
        std::plus{}
    );  // 4 + 16 = 20
}

Example 3: Zip and Enumerate

#include <ranges>
#include <vector>

void zipEnumerate() {
    std::vector<std::string> names = {"Alice", "Bob", "Charlie"};
    std::vector<int> scores = {95, 87, 92};
    
    // Zip two ranges (C++23)
    // for (auto [name, score] : std::views::zip(names, scores)) {
    //     std::cout << name << ": " << score << '\n';
    // }
    
    // Enumerate with indices (C++23)
    // for (auto [idx, name] : std::views::enumerate(names)) {
    //     std::cout << idx << ". " << name << '\n';
    // }
}

Range Adaptors

Create custom views for specific needs:

#include <ranges>

// Custom view: chunk (group elements)
void customAdaptor() {
    std::vector<int> nums = {1, 2, 3, 4, 5, 6, 7, 8, 9};
    
    // chunk is available in C++23
    // auto chunks = nums | std::views::chunk(3);
    // Each chunk: {1,2,3}, {4,5,6}, {7,8,9}
}

Performance Considerations

Performance Benefits:

No intermediate allocations
Lazy evaluation - only compute what's needed
Better optimization opportunities for compilers

#include <ranges>
#include <vector>

void performanceComparison() {
    std::vector<int> data(1'000'000);
    
    // Traditional: creates intermediate vectors
    std::vector<int> temp1, temp2, result;
    std::copy_if(data.begin(), data.end(), std::back_inserter(temp1), 
                 [](int n) { return n % 2 == 0; });
    std::transform(temp1.begin(), temp1.end(), std::back_inserter(temp2),
                   [](int n) { return n * n; });
    std::copy_n(temp2.begin(), 100, std::back_inserter(result));
    
    // Ranges: no intermediate allocations, lazy evaluation
    auto result_view = data
        | std::views::filter([](int n) { return n % 2 == 0; })
        | std::views::transform([](int n) { return n * n; })
        | std::views::take(100);
    
    std::vector<int> result2(result_view.begin(), result_view.end());
}

Common Patterns

Pattern 1: Filter-Transform-Collect

std::vector<int> nums = {1, 2, 3, 4, 5, 6, 7, 8, 9, 10};

auto result = nums
    | std::views::filter([](int n) { return n > 5; })
    | std::views::transform([](int n) { return n * 2; });

std::vector<int> vec(result.begin(), result.end());

Pattern 2: Flat Map (C++23)

std::vector<std::vector<int>> nested = {{1, 2}, {3, 4}, {5, 6}};

// auto flat = nested | std::views::join;  // {1, 2, 3, 4, 5, 6}

Pattern 3: Infinite Sequences

// Generate infinite Fibonacci-like sequence
auto fibonacci = std::views::iota(0)
    | std::views::transform([](int n) { /* compute fib(n) */ return n; })
    | std::views::take(10);

Best Practices

DO:

Use views for lazy, composable operations
Prefer range algorithms over iterator-based ones
Use projections instead of custom comparators
Chain operations with | operator for readability

danger

DON'T:

Store views long-term (they don't own data)
Modify underlying containers while iterating views
Forget to limit infinite ranges with take()
Use views when you need materialized results multiple times

Migration from Traditional STL

// Before (C++17)
std::vector<int> nums = {1, 2, 3, 4, 5};
std::vector<int> evens;
std::copy_if(nums.begin(), nums.end(), std::back_inserter(evens),
             [](int n) { return n % 2 == 0; });
std::transform(evens.begin(), evens.end(), evens.begin(),
               [](int n) { return n * n; });

// After (C++20)
auto result = nums
    | std::views::filter([](int n) { return n % 2 == 0; })
    | std::views::transform([](int n) { return n * n; });

Algorithms - STL algorithms
Iterators - Iterator concepts
Containers - Data structures

Why Ranges?​

Range Concepts​

Views​

Basic Views​

View Composition (Pipelines)​

Common Views​

Range Factories​

Range Algorithms​

Projections​

Practical Examples​

Example 1: Processing Text Lines​

Example 2: Mathematical Operations​

Example 3: Zip and Enumerate​

Range Adaptors​

Performance Considerations​

Common Patterns​

Pattern 1: Filter-Transform-Collect​

Pattern 2: Flat Map (C++23)​

Pattern 3: Infinite Sequences​

Best Practices​

Migration from Traditional STL​

Related Topics​

Why Ranges?

Range Concepts

Views

Basic Views

View Composition (Pipelines)

Common Views

Range Factories

Range Algorithms

Projections

Practical Examples

Example 1: Processing Text Lines

Example 2: Mathematical Operations

Example 3: Zip and Enumerate

Range Adaptors

Performance Considerations

Common Patterns

Pattern 1: Filter-Transform-Collect

Pattern 2: Flat Map (C++23)

Pattern 3: Infinite Sequences

Best Practices

Migration from Traditional STL

Related Topics