Files
C-Plus-Plus/sorting/binary_insertion_sort.cpp
Siddhartha Shankar Padhy d48126b70d Update binary_insertion_sort.cpp
2021-12-07 18:58:41 +05:30

152 lines
5.3 KiB
C++

/**
* \file
* \brief [Binary Insertion Sort Algorithm
* (Insertion Sort)](https://en.wikipedia.org/wiki/Insertion_sort)
*
* \details
* If the cost of comparisons exceeds the cost of swaps, as is the case for example with
* string keys stored by reference or with human interaction (such as choosing one of a pair
* displayed side-by-side), then using binary insertion sort may yield better performance.
* Binary insertion sort employs a binary search to determine the correct location to insert
* new elements, and therefore performs ⌈log2 n⌉ comparisons in the worst case.
* When each element in the array is searched for and inserted this is O(n log n).
* The algorithm as a whole still has a running time of O(n2) on average because of the series * of swaps required for each insertion.
* However it has several advantages such as
* 1. Easy to implement
* 2. For small set of data it is quite efficient
* 3. More efficient that other Quadratic complexity algorithms like
* Selection sort or bubble sort.
* 4. It is efficient to use it when the cost of comparison is high.
* 5. It's stable that is it does not change the relative order of
* elements with equal keys.
* 6. It can sort the array or list as it receives.
*
* Example execution steps:
* 1. Suppose initially we have
* \f{bmatrix}{40 &30 &20 &50 &10\f}
* 2. We start traversing from 40 till we reach 10
* when we reach at 30 we find that it is not at it's correct place so we take 30 and place
* it at a correct position thus the array will become
* \f{bmatrix}{30 &40 &20 &50 &10\f}
* 3. In the next iteration we are at 20 we find that this is also misplaced so
* we place it at the correct sorted position thus the array in this iteration
* becomes
* \f{bmatrix}{20 &30 &40 &50 &10\f}
* 4. We do not do anything with 50 and move on to the next iteration and
* select 10 which is misplaced and place it at correct position. Thus, we have
* \f{bmatrix}{10 &20 &30 &40 &50\f}
*/
#include <algorithm> /// for algorithm functions
#include <cassert> /// for assert
#include <iostream> /// for IO operations
#include <vector> /// for working with vectors
/**
* \namespace sorting
* @brief Sorting algorithms
*/
namespace sorting {
/**
* \brief Binary search function to find the most suitable pace for an element.
* \tparam T The generic data type.
* \param arr The actual vector in which we are searching a suitable place for the element.
* \param val The value for which suitable place is to be found.
* \param low The lower bound of the range we are searching in.
* \param high The upper bound of the range we are searching in.
* \returns the index of most suitable position of val.
*/
template<class T>
int64_t binary_search(std::vector<T> &arr,T val,int64_t low,int64_t high)
{
if (high <= low)
{
return (val > arr[low]) ? (low + 1) : low;
}
int64_t mid = low + (high-low)/2;
if(arr[mid]>val)
{
return binary_search(arr,val,low,mid-1);
}
else if(arr[mid]<val)
{
return binary_search(arr,val,mid+1,high);
}
else
{
return mid+1;
}
}
/**
* \brief Insertion sort function to sort the vector.
* \tparam T The generic data type.
* \param arr The actual vector to sort.
* \returns Void.
*/
template <typename T>
void insertionSort_binsrch(std::vector<T> &arr) {
int64_t n = arr.size();
for (int64_t i = 1; i < n; i++) {
T key = arr[i];
int64_t j = i - 1;
int64_t loc = sorting::binary_search(arr,key,0,j);
while (j >= loc) {
arr[j + 1] = arr[j];
j--;
}
arr[j + 1] = key;
}
}
} // namespace sorting
/**
* @brief Self-test implementations
* @returns void
*/
static void test() {
/* descriptions of the following test */
/* 1st test:
[5, -3, -1, -2, 7] returns [-3, -2, -1, 5, 7] */
std::vector<int64_t> arr1({5, -3, -1, -2, 7});
std::cout << "1st test... ";
sorting::insertionSort_binsrch(arr1);
assert(std::is_sorted(std::begin(arr1), std::end(arr1)));
std::cout << "passed" << std::endl;
/* 2nd test:
[12, 26, 15, 91, 32, 54, 41] returns [12, 15, 26, 32, 41, 54, 91] */
std::vector<int64_t> arr2({12, 26, 15, 91, 32, 54, 41});
std::cout << "2nd test... ";
sorting::insertionSort_binsrch(arr2);
assert(std::is_sorted(std::begin(arr2), std::end(arr2)));
std::cout << "passed" << std::endl;
/* 3rd test:
[7.1, -2.5, -4.0, -2.1, 5.7] returns [-4.0, -2.5, -2.1, 5.7, 7.1] */
std::vector<float> arr3({7.1, -2.5, -4.0, -2.1, 5.7});
std::cout << "3rd test... ";
sorting::insertionSort_binsrch(arr3);
assert(std::is_sorted(std::begin(arr3), std::end(arr3)));
std::cout << "passed" << std::endl;
/* 4th test:
[12.8, -3.7, -20.7, -7.1, 2.2] returns [-20.7, -7.1, -3.7, 2.2, 12.8] */
std::vector<float> arr4({12.8, -3.7, -20.7, -7.1, 2.2});
std::cout << "4th test... ";
sorting::insertionSort_binsrch(arr4);
assert(std::is_sorted(std::begin(arr4), std::end(arr4)));
std::cout << "passed" << std::endl;
}
/**
* @brief Main function
* @return 0 on exit.
*/
int main() {
test(); // run self-test implementations
return 0;
}