Counting Algorithms
An introduction to the 5 main counting algorithms in the C++ standard library: count()
, count_if()
, any_of()
, none_of()
, and all_of()
Let's take a look at some useful counting algorithms available in the standard library. To access these, we need the <algorithm>
header:
#include <algorithm>
In this lesson, we cover the 5 most useful counting algorithms: count()
, count_if()
, any_of()
, none_of()
and all_of()
. Let's get started!
Using count()
The std::ranges::count
algorithm requires two arguments - the range we want to run the algorithm in, and the element we want to count within that range.
Whether or not an element matches what we're looking for is determined by an equality check. Therefore, the type of elements in our collection must implement the equality operator, ==
In this example, we count the number of 4
s in our vector of integers:
#include <algorithm>
#include <iostream>
#include <vector>
int main() {
std::vector Numbers { 1, 2, 3, 4, 4, 5 };
auto Fours { std::ranges::count(Numbers, 4) };
std::cout << "Count of fours: " << Fours;
}
Count of fours: 2
Return Type
The return type of these algorithms can be somewhat cryptic. For count()
, the return type is a template:
std::ranges::range_difference_t
Given we're counting elements in a std::vector<int>
in this case, the type returned from count()
would be even more verbose:
std::ranges::range_difference_t<std::vector<int>> Fours
{
std::ranges::count(Numbers, 4)
};
It's common to just use auto
in this scenario, as we did in the original example. Alternatively, it's reasonable to implicitly cast the return value to something simpler, like a 64-bit integer or size_t
:
int64_t Fours { std::ranges::count(Numbers, 4) };
Using count_if()
The basic std::ranges::count()
algorithm uses the equality operator ==
to determine if an element in our container matches the object we're looking for.
The std::ranges::count_if()
algorithm lets us change this behavior. Rather than using the ==
operator to determine if two objects are equal, count_if()
uses a function, which we provide as an argument.
The function will be called for every object in the collection, receiving that object as an argument. Our function should return true
if we want this object to be included in the count, and false
otherwise. A function that returns a boolean is also known as a predicate.
Below, we have a predicate that returns true
if a provided argument is even. We pass this predicate to std::ranges::count_if()
to count the number of elements in our range that are even:
#include <algorithm>
#include <iostream>
#include <vector>
int main() {
std::vector Numbers { 1, 2, 3, 4, 4, 5 };
auto isEven{[](int x) { return x % 2 == 0; }};
auto EvenCount {
std::ranges::count_if(Numbers, isEven)
};
std::cout << "Even Count: " << EvenCount;
}
Even Count: 3
Using any_of()
The std::ranges::any_of()
algorithm also accepts a predicate function. The algorithm will return true
if our predicate returns true
for any element in the range.
In the following example, we return true
if any number in our collection is even:
#include <algorithm>
#include <iostream>
#include <vector>
int main() {
std::vector Numbers { 1, 2, 3, 4, 4, 5 };
auto isEven{[](int x) { return x % 2 == 0; } };
bool IncludesEven {
std::ranges::any_of(Numbers, isEven)};
std::cout << "An even number "
<< (IncludesEven ? "is" : "is not")
<< " included";
}
An even number is included
Using none_of()
The std::ranges::none_of()
algorithm will run a predicate on everything in our collection, and return true
if every predicate returns false
.
If our predicate returns true
for any element in the collection, the algorithm will stop evaluating further elements, and return false
.
#include <algorithm>
#include <iostream>
#include <vector>
int main() {
std::vector Numbers { 1, 2, 3, 4, 4, 5 };
auto isEven{ [](int x) { return x % 2 == 0; } };
bool NoEvenNumbers {
std::ranges::none_of(Numbers, isEven)};
std::cout << "The range contains "
<< (NoEvenNumbers ? "no" : "some")
<< " even numbers";
}
The range contains some even numbers
This algorithm will return true
if the range has no elements:
#include <algorithm>
#include <iostream>
#include <vector>
int main() {
std::vector<int> Numbers { };
auto isEven { [](int x) { return x % 2 == 0; } };
bool Result {
std::ranges::none_of(Numbers, isEven)};
std::cout << (Result ? "true" : "false");
}
true
Using all_of()
The std::ranges::all_of()
algorithm will run a predicate on everything in our collection, and return true
if every invocation of the predicate returns true
.
If our predicate returns false
for any element in the collection, the algorithm will stop evaluating further elements, and return false
.
#include <algorithm>
#include <iostream>
#include <vector>
int main() {
std::vector Numbers { 1, 2, 3, 4, 4, 5 };
auto isEven{[](int x) { return x % 2 == 0; }};
bool AllEvenNumbers {
std::ranges::all_of(Numbers, isEven)};
std::cout << "The range is "
<< (AllEvenNumbers ? "all" : "NOT all")
<< " even numbers";
}
The range is NOT all even numbers
Note, that this algorithm will return true
if the range has no elements:
#include <algorithm>
#include <iostream>
#include <vector>
int main() {
std::vector<int> Numbers{};
auto isEven{[](int x) { return x % 2 == 0; }};
bool Result{
std::ranges::all_of(Numbers, isEven)};
std::cout << Result ? "true" : "false";
}
true
Using Member Functions as Predicates
Typically when using these algorithms, the predicates we need will be class methods of the objects contained within our range. When that is the case, we generally don't need to define a predicate function. We can just pass a reference to the class method:
#include <algorithm>
#include <iostream>
#include <vector>
class Player {/*...*/};
int main() {
std::vector Party {
Player{"Roderick", 100},
Player{"Anna", 200},
Player{"Robert", 500}
};
bool EveryoneAlive{std::ranges::all_of(
Party, &Player::isAlive)};
std::cout << "Everyone "
<< (EveryoneAlive ? "is" : "is not")
<< " alive";
}
Everyone is alive
Projection
All the algorithms in this lesson accept an additional optional argument, which can be a projection function. We cover projection in detail here:
Projection Functions
Learn how to use projection functions to apply range-based algorithms on derived data
In this example, we project the objects in our container to their absolute value before running the count()
algorithm. Two of those projections are equal to 1
, so our algorithm returns 2
:
#include <algorithm>
#include <iostream>
#include <vector>
int Projector(int x) { return std::abs(x); }
int main() {
std::vector Nums{-2, -1, 0, 1, 2};
auto Count {std::ranges::count(
Nums, 1, Projector)};
std::cout << "Count: " << Count;
}
Count: 2
In this example, we use Player::GetName()
as our projection function, projecting our Player
objects to a std::string
before comparing them to the "Anna"
string:
#include <algorithm>
#include <iostream>
#include <vector>
class Player {/*...*/};
int main() {
std::vector Party {
Player{"Roderick", 100},
Player{"Anna", 200},
Player{"Robert", 500}
};
auto AnnaCount{std::ranges::count(
Party, "Anna", &Player::GetName)};
std::cout << "AnnaCount: " << AnnaCount;
}
AnnaCount: 1
Using Iterators
As we covered in the introduction to range-based algorithms, std::ranges::count()
and all the other algorithms we covered in this lesson allow us to define our range as an iterator-sentinel pair.
Below, we use this technique to exclude the first and last objects from our count:
#include <algorithm>
#include <iostream>
#include <vector>
int main() {
std::vector Numbers{1, 2, 3, 4, 4};
auto Fours{std::ranges::count(
Numbers.begin() + 1, Numbers.end() - 1, 4)};
std::cout << "Count of fours: " << Fours;
}
Count of fours: 1
The concept of a range was introduced in C++20. When targeting older specifications, we can use alternative versions of these algorithms that work directly with iterators instead.
These are available by omitting the ranges
qualification from the identifiers. For example, the iterator variant of std::ranges::count()
is available as std::count()
:
#include <algorithm>
#include <iostream>
#include <vector>
int main() {
std::vector Numbers{1, 2, 3, 4, 4, 5};
auto Fours{std::count(
Numbers.begin(), Numbers.end(), 4)};
std::cout << "Count of fours: " << Fours;
}
Count of fours: 2
Summary
In this lesson, we explored the counting algorithms that are part of the C++20 std::ranges
library, and available by including the <algorithm>
header.
Key Takeaways
- Understanding of
std::ranges::count()
for counting specific elements in a range. - Use of
std::ranges::count_if()
to count elements that meet a predicate condition. - Application of
std::ranges::any_of()
,none_of()
, andall_of()
for conditional checks across elements. - Insight into the performance benefits of using specific algorithms like
any_of()
over alternatives likecount_if()
in certain scenarios. - Using predicates in customizing the behavior of counting algorithms.
- How to use member functions as predicates for more complex conditions.
- The role of projection functions in transforming elements before applying counting algorithms.
- Review of using iterator-sentinel pairs for defining custom ranges.
Minimum and Maximum Algorithms
An introduction to the seven minimum and maximum algorithms in the C++ standard library: clamp()
, min()
, min_element()
, max()
, max_element()
, minmax()
, and minmax_element()
.