Debugging a merge sort - c++

void CensusData::mergeSort(int type) {
if(type == 0)
MERGE_SORT(type, 0, data.size());
}
void CensusData::MERGE_SORT(int type, int p, int r){
//int q;
//cout << "data size " << data.size() << endl;
std::cout << "MERGE_SORT START ///("<< p << ", " << r << ")" <<std::endl;
if(p < r)
{
int q = (p + r)/2;
MERGE_SORT(type, p, q);
MERGE_SORT(type, q + 1, r);
MERGE(type, p, q ,r);
}
}
void CensusData::MERGE(int type, int p, int q, int r){
if(type == 0)
{
std::cout << "MERGING" << std::endl;
//int n1;
//int n2;
int n1 = q - p + 1;
int n2 = r - q;
int L[n1 + 1];
int R[n2 + 1];
for(int i = 1; i < n1; i++)
{
cout << "filling Left Array" << endl;
L[i] = data[p + i - 1]->population;
}
for(int j = 1; j < n2; j++)
{
cout << "filling Right Array" << endl;
R[j] = data[q + j]->population;
}
int i = 1;
int j = 1;
for(int k = p; p < r; p++)
{
cout << "for loop: " << endl;
if(L[i] <= R[j])
{
cout << "TRUE" << endl;
data[k]->population = L[j];
i = i + 1;
}
/*else if(data[k]->population == R[j])
{
cout << "FALSE" << endl;
j = j + 1;
}*/
else
{
data[k]->population = R[j];
j = j + 1;
}
}
}
}
do not worry about type, it wont effect this program at all. basically i am trying to make a merge sort that will take a vector containing an integer, the vector looks like this:
class Record { // declaration of a Record
public:
std::string* city;
std::string* state;
int population;
Record(std::string&, std::string&, int);
~Record();
};
std::vector<Record*> data;
basically i have been trying to get it to actually sort, but it doesn't seem to work at all, i have even seen garbage in the program.
example input:
237 812826 68642
output:
4484540 812826 68642
Note: all of the rest of the program works fine (tested it with an insertion sort) only this part is not working.

Take a look at lecture 15 of the excellent Stanford Universities course Programming Abstractions. It covers all kinds of sorts including merge:
http://see.stanford.edu/see/lecturelist.aspx?coll=11f4f422-5670-4b4c-889c-008262e09e4e
You can even get the source code from SourceForge:
http://sourceforge.net/projects/progabstrlib/files/

Related

Program containing threading in cpp is not executed completely

Code given below is not executed completely;
I have looked for everything on web but I don't know why it is working for starting numbers from nums (i.e. 1000 and sometimes 5000) and after that it starts execution but in between program terminates itself and stopes working.
#include <bits/stdc++.h>
// #include <iostream>
// #include <chrono>
// #include <vector>
#define UPPER_LIMIT 10
using namespace std;
using namespace std::chrono;
bool inTimeLimit = true;
bool isFinished = false;
bool isRunning = true;
class Timer {
public:
time_point<high_resolution_clock> start, end;
Timer() {
start = high_resolution_clock::now();
}
~Timer() {
end = high_resolution_clock::now();
auto durationTime = durationCounter();
cout << "\n\nTaken time Duration " << (unsigned long long)durationTime << " us; " << (unsigned long long)durationTime * 0.001 << "ms.";
}
float durationCounter() {
auto currTime = high_resolution_clock::now();
auto durationTime = duration_cast<microseconds>(currTime - start);
return durationTime.count();
}
};
void printVector(vector <int> v) {
cout << endl;
for (int x : v) {
cout << setw(3) << x << " ";
}
}
void printVectorToFile(ofstream &fout , vector <int> v, string msg) {
fout << "\n\n===================\n\n";
fout << msg << endl;
fout << endl;
for (int x : v) {
fout << setw(5) << x << " ";
}
fout << endl;
}
void swap (int *a, int *b) {
int temp = *a;
*a = *b;
*b = temp;
}
vector <int> randomArrayGenerator(int n) {
vector<int> v(n);
for (int i = 0; i < n; ++i)
v[i] = i + 1;
srand(time(0));
for (int i = 0; i < n; ++i)
{
int pos = rand() % n;
swap(&v[i], &v[pos]);
}
return v;
}
string sortingChecker(vector<int> v) {
for (int i = 0; i < (int)v.size() - 1; ++i)
{
if (v[i] > v[i + 1]) return "false";
}
return "true";
}
bool sortChecker(vector<int> v) {
for (int i = 0; i < (int)v.size() - 1; ++i)
{
if (v[i] > v[i + 1]) return false;
}
return true;
}
// Merge function
void merge(vector <int> &v, int begin, int middle, int end) {
vector <int> left, right;
for (int i = begin; i < middle + 1; ++i)
{
left.push_back(v[i]);
}
for (int i = middle + 1; i <= end; ++i)
{
right.push_back(v[i]);
}
int p1 = 0, p2 = 0, n1 = left.size(), n2 = right.size(), p = begin;
while ((p1 < n1 ) || (p2 < n2)) {
if ((p1 != n1 ) && ((p2 == n2) || left[p1] < right[p2]))
v[p++] = left[p1++];
else
v[p++] = right[p2++];
}
}
void mergeSortByIteration(vector <int> &v, bool &isTimeDelayed) {
int low = 0, high = v.size();
cout << "Thread ID: " << this_thread::get_id() << endl;
// n :for taking individual block of vector containing number of elements n=[1,2,4,8,..]
for (int n = 1; n < high; n *= 2) {
if (isTimeDelayed) return;
// taking block according to n and then sorting them by merge function
// n=1 => i=0,2,4,8,16
// n=2 => i=0,4,8
for (int i = 0; i < high; i += 2 * n) {
if (isTimeDelayed) return;
int begin = i;
int mid = i + n - 1;
int end = min(i + 2 * n - 1 , high - 1);
merge(v, begin, mid, end);
}
}
}
// Merge by recurision
void mergeSortByRecursion (vector <int> &v, int begin, int end, bool &isTimeDelayed) {
if (end <= begin || isTimeDelayed) return;
int middle = begin + (end - begin) / 2;
mergeSortByRecursion(v, begin, middle, isTimeDelayed);
mergeSortByRecursion(v, middle + 1, end, isTimeDelayed);
merge(v, begin, middle, end);
}
int main() {
int nums[] = {1000, 5000, 10000, 50000, 100000};
// int nums[] = {50000};
ofstream vectorOutput ;
vectorOutput.open("outputTexts\\prac1_resultedArrays.txt", ios::trunc);;
for (int n : nums)
// ``````` Merge by Iteration ````````
{
vector<int> num, arr = randomArrayGenerator(n);
cout << "\n=======";
cout << "\n\nMerge by Iteration:" << endl;
num = arr;
cout << "Array size: " << num.size() << endl;
bool isTimeOut = false, isSorted = false;
Timer timer;
std::thread worker(mergeSortByIteration, ref(num), ref(isTimeOut));
// mergeSortByIteration(num, isTimeOut);
// std::thread worker(mergeSortByRecursion, ref(num), 0, n - 1, ref(isTimeOut));
while ( ( ( timer.durationCounter() / 1000000 ) < 5) && (!isSorted ) ) {
// this_thread::sleep_for(seconds(1));
// cout << timer.durationCounter() << " ";
isSorted = sortChecker(num);
}
if ( ( ( ( timer.durationCounter() / 1000000 ) > 5) && (!isSorted ) ) )
{
isTimeOut = true;
cout << endl << "!!!!!Execution Terminated ---- Time Limit reached!!!!!!" << endl;
}
if (worker.joinable())
worker.join();
printVector(num);
cout << "\nCheck result for sorted Vector:" << (isSorted ? "true" : "false") << endl;
// printVectorToFile(vectorOutput, num, "Merge By Iteration for size:" + to_string(n) );
}
cout << "\n\ndone" << endl;
return 0;
}
can anyone help me out here?
If issue is not clear fill free to ask.

How to pass an array to a class?

#include <iostream>
#include <string>
using namespace std;
class Term;
class Polynomial;
class Term {
public:
friend class Polynomial;
void set_term(string s);
int get_pow();
int get_coefficient();
private:
int coefficient;
int pow;
};
class Polynomial {
public:
friend class Term;
Polynomial(int s, Term t[]);
int get_size();
Term *myterm;
private:
int P_size;
};
void Term::set_term(string s) {
coefficient = stoi(s.substr(1, 1));
pow = stoi(s.substr(4, 1));
if (s[0] == '-') {
coefficient = -coefficient;
}
}
int Term::get_coefficient() { return coefficient; }
int Term::get_pow() { return pow; }
Polynomial::Polynomial(int s, Term t[]) {
P_size = s;
myterm = new Term[s];
for (int i = 0; i < s; i++) {
myterm[i].coefficient = t[i].coefficient;
cout << i << " Term " << t[i].coefficient << endl;
cout << i << " Polynomial " << myterm[i].coefficient << endl;
myterm[i].pow = t[i].pow;
}
}
Polynomial::get_size() { return P_size; }
int main() {
string x1;
cin >> x1;
int size_x1 = x1.size();
Term term1[size_x1];
for (int i = 0; i < size_x1; i += 5) {
term1[i].set_term(x1.substr(i, 5));
}
Polynomial p1(size_x1 / 5, term1);
for (int i = 0; i < size_x1; i += 5) {
cout << term1[i].get_coefficient() << "x^";
cout << term1[i].get_pow() << endl;
}
cout << "------------------" << endl;
for (int i = 0; i < p1.get_size(); i++) {
if (p1.myterm[i].get_coefficient() > 0)
cout << "+";
cout << p1.myterm[i].get_coefficient();
cout << "x^";
cout << p1.myterm[i].get_pow() << endl;
}
return 0;
}
term1 in main is working right but when I pass it to p1(Polynomial) just t[0] and, myterm[0]
is true I mean I pass term1 to p1 as t so t[0] = term1[0] and myterm[0] = t[0]
Polynomial::Polynomial(int s, Term t[]) {
P_size = s;
myterm = new Term[s];
for (int i = 0; i < s; i++) {
myterm[i].coefficient = t[i].coefficient;
cout << i << " Term " << t[i].coefficient << endl;
cout << i << " Polynomial " << myterm[i].coefficient << endl;
myterm[i].pow = t[i].pow;
}
}
in Polynomial constructor both t[1].coeffient and, myterm[1].coefficient are 4941660(for instance)
however, t is term1 and in term1 term1[1] is the 4(for instance)
if term1[1] is 4 so t[1] should be 4 but it's not so myterm[1], is not 4 too.
There's a lot of confusion over sizes here
string x1;
cin >> x1;
int size_x1 = x1.size();
Term term1[size_x1];
for (int i = 0; i < size_x1; i += 5) {
term1[i].set_term(x1.substr(i, 5));
}
Polynomial p1(size_x1 / 5, term1);
You didn't say what your input is, but suppose it is "1234567890" then size_x1 equals 10, so term1 will also have size 10. Then your for loop will execute twice, with i equal to 0 and i equal to 5, so you will assign 12345 to term1[0] and 67890 to term1[5]. That makes no sense to me. term1 has size 10 but the only elements assigned to are zero and five.
Then you create p1 with size size_x1/5 which equals 2. So the constructor for p1 will look at term1[0] and term1[1], but in the for loop you assigned to term1[0] and term1[5].
You have to think carefully about what the code you are writing is actually doing. The compiler will do exactly what you tell it to. It's not like talking to a person who can understand what you really meant. So you have to be very careful. This is I think you really meant (but I could be wrong)
string x1;
cin >> x1;
int size_x1 = x1.size();
int size_term1 = size_x1/5;
Term term1[size_term1];
for (int i = 0; i < size_term1 ; ++i) {
term1[i].set_term(x1.substr(5*i, 5));
}
Polynomial p1(size_term1, term1);
I created a new variable size_term1 with the size of the term array, which is now the sizeof the input divided by 5. And I changed the for loop so that it assigns to term1[0] and term1[1].
So your problem has nothing to do with passing an array to a class, which you did perfectly correctly. It was not thinking clearly about the code you were writing, and making mistakes in the logic.

Difference in runtime C++ using different data types

Ok, so I was doing a tiny project for school and I can't find the answer anywhere to why this small change in code makes it finish in no time when number m gets higher. Look at the variable "k" I change it from int to long.
I'm trying to find the longest sequence in the Collatz sequence between 1 and 1000000
void lengstaRuna() {
cout << "Hæsta tala?:";
int m;
cin >> m;
int lengstaRuna = 0;
int talaLengstuRunu = 0;
int k;
for(int i = 2; i < m; i++) {
int lengd = 1;
k = i;
while(k != 1) {
if(k % 2 == 0) {
k = k/2;
} else {
k = k*3 +1;
}
lengd++;
}
if(lengd > lengstaRuna) {
lengstaRuna = lengd;
talaLengstuRunu = i;
}
}
cout << "Lengsta runa: " << lengstaRuna << endl;
cout << "Tala lengstu runu: " << talaLengstuRunu << endl;
}
void lengstaRuna() {
cout << "Hæsta tala?:";
int m;
cin >> m;
int lengstaRuna = 0;
int talaLengstuRunu = 0;
long k;
for(int i = 2; i < m; i++) {
int lengd = 1;
k = i;
while(k != 1) {
if(k % 2 == 0) {
k = k/2;
} else {
k = k*3 +1;
}
lengd++;
}
if(lengd > lengstaRuna) {
lengstaRuna = lengd;
talaLengstuRunu = i;
}
}
cout << "Lengsta runa: " << lengstaRuna << endl;
cout << "Tala lengstu runu: " << talaLengstuRunu << endl;
}
The question is simple: Why does it run so much faster when input m==1000000?
I see what's happening here. Basically, above certain value for your input, the int is overflowing since you are doing k*3.
I modified your code to check this (see below). Upto input value of around 113000, the max your 'k' has to hold is 1570824735 (close to INT_MAX 2147483647). Anything 114000 or above, 'k' overflows and the code goes into uncharted territory. That problem doesn't happen when you use long of course.
./a.out 113000
j: 1570824735
Lengsta runa: 354
Tala lengstu runu: 106239
#include <iostream>
#include <string>
using namespace std;
void lengstaRuna(int m) {
int lengstaRuna = 0;
int talaLengstuRunu = 0;
int k;
long j = 0;
for(int i = 2; i < m; i++) {
int lengd = 1;
k = i;
while(k != 1) {
if(k % 2 == 0) {
k = k/2;
} else {
if (k*3 > j)
j = k*3;
k = k*3 +1;
}
lengd++;
}
if(lengd > lengstaRuna) {
lengstaRuna = lengd;
talaLengstuRunu = i;
}
}
cout << "j: " << j << endl;
cout << "Lengsta runa: " << lengstaRuna << endl;
cout << "Tala lengstu runu: " << talaLengstuRunu << endl;
}
int main (int ac, char** av) {
std::string::size_type sz;
lengstaRuna(std::stoi(av[1]));
}

why increment variable changing the value of the array when they have different names

Can someone please help me. I am struggling to find in my code why the last value in column B always gets incremented by one. I have written some code since its an assignment due today. I also cant figure out why the last value in column B is not equal to 196 because in the reset function it sets all the values in the array to 196 . Any suggestion would be appreciated. Thank you in advance
#include <iostream> //includes cin cout
#include <iomanip>
using namespace std; //setting up the environment
const int NUMBER_OF_ROWS = 3;
const int NUMBER_OF_COLUMNS = 3;
void printAllSeats(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]);
void reset(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]);
void askForUsersSeat(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS], int &SeatCountNumber, bool &anyFreeSeats);
bool isFull(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]);
bool isEmpty(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]);
int main() { //main starts
int maxSeats;
int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS];
int SeatCountNumber = 0;
bool anyFreeSeats;
reset(seats);
anyFreeSeats = true;
SeatCountNumber = 0;
while (anyFreeSeats) {
printAllSeats(seats);
askForUsersSeat(seats, SeatCountNumber, anyFreeSeats);
}
system("pause");
return 0;
} //main ends
void printAllSeats(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]) {
cout << endl;
cout << setw(10) << " - = Available R = Reserved\n\n";
for (int i = 0; i <= NUMBER_OF_ROWS; i++) {
cout << setw(15) << i << " ";
for (int j = 0; j < NUMBER_OF_COLUMNS; j++) {
if (i == 0) {
cout << " " << static_cast<char>(j + 65) << " ";
} else {
cout << " " << static_cast<char>(seats[i][j]) << " ";
}
}
cout << endl;
}
cout << endl;
}
void reset(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]) {
//set all values in array to 196
for (int i = 0; i <= NUMBER_OF_ROWS; i++) {
for (int j = 0; j <= NUMBER_OF_COLUMNS; j++) {
seats[i][j] = 196;
}
}
}
void askForUsersSeat(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS], int &SeatCountNumber, bool &anyFreeSeats) {
int seatChoiceNumber;
char seatChoiceLetter;
int letter;
int maxSeats = NUMBER_OF_ROWS * NUMBER_OF_COLUMNS;
cout << "Seat (Row, Column): ";
cin >> seatChoiceNumber >> seatChoiceLetter;
letter = static_cast<int>(toupper(seatChoiceLetter));
if (((letter >= 65) && (letter < (65 + NUMBER_OF_COLUMNS))) && ((seatChoiceNumber > 0) && (seatChoiceNumber <= NUMBER_OF_ROWS))) {
if (seats[(seatChoiceNumber)][(letter - 65)] == 82) {
} else {
seats[(seatChoiceNumber)][(letter - 65)] = 82;
SeatCountNumber++; //this changes last value in column B for some reason
if (SeatCountNumber < maxSeats) {
anyFreeSeats = true;
}
else if (SeatCountNumber > maxSeats) {
printAllSeats(seats);
anyFreeSeats = false;
}
}
} else {
}
}
I kind of cleaned up the code a bit. It seems you found your answer in the comments, so I just did some indentation. Try and eliminate whitespaces in your code (mind you, the one I am putting here is not perfect either, but you get the point). Clean and easy to read code doesn't only make it better for you, but as you get higher up in the industry and other people begin reading and working on your code, having clean and easy to read code really helps :)
#include <iostream> //includes cin cout
#include <iomanip>
using namespace std; //setting up the environment
const int NUMBER_OF_ROWS = 3;
const int NUMBER_OF_COLUMNS = 3;
void printAllSeats(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]);
void reset(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]);
void askForUsersSeat(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS], int &SeatCountNumber, bool &anyFreeSeats);
bool isFull(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]);
bool isEmpty(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS]);
int main()
{
int maxSeats;
int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS];
int SeatCountNumber = 0;
bool anyFreeSeats;
reset(seats);
anyFreeSeats = true;
SeatCountNumber = 0;
while (anyFreeSeats)
{
printAllSeats(seats);
askForUsersSeat(seats, SeatCountNumber, anyFreeSeats);
}
system("pause");
return 0;
} //main ends
void printAllSeats(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS])
{
cout << endl;
cout << setw(10) << " - = Available R = Reserved\n\n";
for (int i = 0; i <= NUMBER_OF_ROWS; i++)
{
cout << setw(15) << i << " ";
for (int j = 0; j < NUMBER_OF_COLUMNS; j++)
{
if (i == 0)
{
cout << " " << static_cast<char>(j + 65) << " ";
}
else
{
cout << " " << static_cast<char>(seats[i][j]) << " ";
}
}
cout << endl;
}
cout << endl;
}
void reset(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS])
{
//set all values in array to 196
for (int i = 0; i <= NUMBER_OF_ROWS; i++)
{
for (int j = 0; j <= NUMBER_OF_COLUMNS; j++)
{
seats[i][j] = 196;
}
}
}
void askForUsersSeat(int seats[NUMBER_OF_ROWS][NUMBER_OF_COLUMNS], int &SeatCountNumber, bool &anyFreeSeats)
{
int seatChoiceNumber;
char seatChoiceLetter;
int letter;
int maxSeats = NUMBER_OF_ROWS * NUMBER_OF_COLUMNS;
cout << "Seat (Row, Column): ";
cin >> seatChoiceNumber >> seatChoiceLetter;
letter = static_cast<int>(toupper(seatChoiceLetter));
if (((letter >= 65) && (letter < (65 + NUMBER_OF_COLUMNS))) && ((seatChoiceNumber > 0) && (seatChoiceNumber <= NUMBER_OF_ROWS)))
{
if (seats[(seatChoiceNumber)][(letter - 65)] == 82)
{
}
else
{
seats[(seatChoiceNumber)][(letter - 65)] = 82;
SeatCountNumber++; //this changes last value in column B for some reason
if (SeatCountNumber < maxSeats)
{
anyFreeSeats = true;
}
else if (SeatCountNumber > maxSeats)
{
printAllSeats(seats);
anyFreeSeats = false;
}
}
}
else {
}
}
Note: Some more whitespaces could even come out but I generally like to have spaces after certain statements (personal preference).

Merge sort code debugging

I am trying to write a code for merge sort. I am not getting the correct output. I am following this pseudocode link Following is my code. I pass my unsorted array into merge_sort function and call merge function recursively to sort and combine the sub arrays.I know there are more simpler and efficient ways to write code for merge sort but I want to try on my own otherwise I won't learn. Thanks in advance.
int* merge_sort(int* a,int size)
{
//cout<<size;
//cout<<"hi";
if(size == 1)
{
//cout<<"less";
//cout<<a[0];
return a;
}
int* left;
int* right;
int middle = ceil(size/2);
left = new int(middle);
right = new int(middle);
for(int i=0;i<middle;i++)
{
left[i]=a[i];
//cout<<left[i];
}
cout<<"\t";
for(int j=middle;j<size;j++)
{
right[j]=a[j];
//cout<<right[j];
}
cout<<"\t";
left = merge_sort(left,middle);
//if(size==2)
//cout<<left[0];
right = merge_sort(right,middle);
//if(size==2)
//cout<<right[0];
return merge(left,right,middle);
}
int* merge(int* l,int* r,int m)
{
int* result;
result = new int(2*m); //to store the output
int lsize=m; // to keep track of left sub list
int rsize=m; // to keep track of right sub list
int counter = 0; // will use to index result
//cout<<m;
while(lsize>0 || rsize>0)
{
if(lsize>0 && rsize>0)
{
if(l[0]<=r[0])
{
result[counter]=l[0];
counter++; //to store next value in result
lsize--;
l=&l[1]; //decrementing the size of left array
}
else
{
result[counter]=r[0];
counter++;
rsize--;
r=&r[1]; //dec. size of right array
}
}
else if(lsize>0)
{
result[counter]=l[0];
counter++;
lsize--;
l=&l[1];
}
else if(rsize>0)
{
result[counter]=l[0];
counter++;
lsize--;
l=&l[1];
}
}
return result;
}
Your code:
int *left = new int(middle);
allocates a single integer initialized to middle. You need:
int *left = new int [middle];
which allocates an array of middle integers. Rinse and repeat for int *right. Actually, you need to use:
int *right = new int [size - middle];
This gets the correct size for the right array. You then have to modify the recursive call to merge_sort() for the right sub-array:
merge_sort(right, size - middle);
Finally, you have to rewrite merge() to take the size of the left array and the size of the right array independently, because they may be of different sizes. For example, if you sort 10 elements,
you then end up with a call to merge two arrays of 5 (which is fine), but at the next level you need to merge an array of 2 and an array of 3 elements (and you're hosed).
The allocation of result also has the () vs [] allocation problem. And there are some other as yet unresolved problems. But these are important steps in the right direction.
As mentioned in a comment to the question, you have a monumental memory leakage problem, too. What's more, it is not trivial to fix because merge_sort() does an early exit without allocating new memory, so it isn't as simple as 'delete the memory returned by merge_sort()'.
Copy and paste is wonderful until you forget to edit the pasted copy correctly:
else if (lsize > 0)
{
result[counter] = l[0];
counter++;
lsize--;
l = &l[1];
}
else if (rsize > 0)
{
result[counter] = l[0];
counter++;
lsize--;
l = &l[1];
}
Methinks you should be using r and rsize in the second of these blocks.
This still isn't the whole story...
And the residual problem (apart from memory management, which is still 100% leaky and problematic) is:
for(int j=middle;j<size;j++)
{
right[j]=a[j];
//cout<<right[j];
}
You're copying into parts of right that you've not allocated. You need something more like:
for(int j = 0; j < size - middle; j++)
{
right[j] = a[j + middle];
//cout<<right[j];
}
This code works as long as you always sort at least two items at the top level (you crash freeing unallocated space if you sort 1 item — that's part of the memory management problem).
#include <iostream>
using namespace std;
namespace {
int *merge(int *l, int m, int *r, int n);
void dump_array(int *a, int size)
{
int i;
cout << size << ": ";
for (i = 0; i < size; i++)
{
cout << ' ' << a[i];
if (i % 10 == 9)
cout << '\n';
}
if (i % 10 != 0)
cout << '\n';
}
};
int *merge_sort(int *a, int size)
{
cout << "-->> merge_sort:\n";
dump_array(a, size);
if (size <= 1)
{
cout << "<<-- merge_sort: early return\n";
return a;
}
int middle = size/2;
int *left = new int[middle];
int *right = new int[size - middle];
cout << middle << ": ";
for (int i = 0; i < middle; i++)
{
left[i] = a[i];
cout << ' ' << left[i];
}
cout << "\n";
cout << (size - middle) << ": ";
for (int j = 0; j < size - middle; j++)
{
right[j] = a[j + middle];
cout << ' ' << right[j];
}
cout << "\n";
cout << "MSL:\n";
int *nleft = merge_sort(left, middle);
cout << "NL: ";
dump_array(nleft, middle);
cout << "OL: ";
dump_array(left, middle);
cout << "OR: ";
dump_array(right, size - middle);
cout << "MSR:\n";
int *nright = merge_sort(right, size - middle);
cout << "NR: ";
dump_array(nright, size - middle);
cout << "NL: ";
dump_array(nleft, middle);
cout << "OL: ";
dump_array(left, middle);
cout << "OR: ";
dump_array(right, size - middle);
int *result = merge(nleft, middle, nright, size - middle);
cout << "<<-- merge_sort:\n";
dump_array(result, size);
return result;
}
namespace {
int *merge(int *l, int m, int *r, int n)
{
int *result = new int[m + n];
int lsize = m;
int rsize = n;
int counter = 0;
cout << "-->> merge: (" << m << "," << n << ")\n";
dump_array(l, m);
dump_array(r, n);
while (lsize > 0 || rsize > 0)
{
if (lsize > 0 && rsize > 0)
{
if (l[0] <= r[0])
{
result[counter] = l[0];
cout << "C: " << counter << "; L = " << l[0] << "; LS = " << lsize << '\n';
counter++;
lsize--;
l++;
}
else
{
result[counter] = r[0];
cout << "C: " << counter << "; R = " << r[0] << "; RS = " << rsize << '\n';
counter++;
rsize--;
r++;
}
}
else if (lsize > 0)
{
result[counter] = l[0];
cout << "C: " << counter << "; L = " << l[0] << "; LS = " << lsize << '\n';
counter++;
lsize--;
l++;
}
else if (rsize > 0)
{
result[counter] = r[0];
cout << "C: " << counter << "; R = " << r[0] << "; RS = " << rsize << '\n';
counter++;
rsize--;
r++;
}
}
cout << "<<-- merge:\n";
dump_array(result, m+n);
return result;
}
};
int main()
{
for (int i = 2; i <= 10; i++)
{
int array1[] = { 9, 3, 5, 7, 1, 8, 0, 6, 2, 4 };
cout << "\nMerge array of size " << i << "\n\n";
int *result = merge_sort(array1, i);
delete[] result;
}
return 0;
}
This is the debug-laden code. It's the level to which I went to get the result. I could perhaps have used a debugger. Were I on a machine where valgrind works, it might have helped too (but it does not work on Mac OS X 10.8.x, sadly).
There are still many, many ways to improve the code — including the memory management. You'd probably find it easiest to pass the input array to merge() for use as the result array (avoiding the memory allocation in that code). This would reduce the memory management burden.
When you remove the debug code, you'll need to call the dump_array() function in the main() program to get the before and after sorting array images.
Code converted to template functions and leak-free
I've simplified the code a fair bit, especially in the merge() function. Also, more as a matter of curiosity than anything else, converted it to a set of template functions, and then used them with 4 different array types (int, double, std::string, char). The amount of debugging has been dramatically reduced, and the main debugging is conditional on being compiled with -DTRACE_ENABLED now.
The code is now leak-free; valgrind on a Linux box (virtual machine) gives it a clean bill of health when there are no exceptions. It is not guaranteed exception-safe, though. In fact, given the naked uses of new and delete, it is pretty much guaranteed not to be exception-safe. I've left the namespace control in place, but I'm far from convinced it is really correct — indeed, I'd lay odds on it not being good. (I'm also curious if anyone has any views on how to layout code within a namespace { … }; block; it seems odd not indenting everything inside a set of braces, but …)
#include <iostream>
using namespace std;
namespace {
#if !defined(TRACE_ENABLED)
#define TRACE_ENABLED 0
#endif
enum { ENABLE_TRACE = TRACE_ENABLED };
template <typename T>
void merge(T *l, int m, T *r, int n, T *result);
template <typename T>
void dump_array(const char *tag, T *a, int size)
{
int i;
cout << tag << ": (" << size << ") ";
for (i = 0; i < size; i++)
{
cout << " " << a[i];
if (i % 10 == 9)
cout << '\n';
}
if (i % 10 != 0)
cout << '\n';
}
};
template <typename T>
void merge_sort(T *a, int size)
{
if (size <= 1)
return;
if (ENABLE_TRACE)
dump_array("-->> merge_sort", a, size);
int middle = size/2;
T *left = new T[middle];
T *right = new T[size - middle];
for (int i = 0; i < middle; i++)
left[i] = a[i];
for (int j = 0; j < size - middle; j++)
right[j] = a[j + middle];
merge_sort(left, middle);
merge_sort(right, size - middle);
merge(left, middle, right, size - middle, a);
delete [] left;
delete [] right;
if (ENABLE_TRACE)
dump_array("<<-- merge_sort", a, size);
}
namespace {
template <typename T>
void merge(T *l, int m, T *r, int n, T *result)
{
T *l_end = l + m;
T *r_end = r + n;
T *out = result;
if (ENABLE_TRACE)
{
cout << "-->> merge: (" << m << "," << n << ")\n";
dump_array("L", l, m);
dump_array("R", r, n);
}
while (l < l_end && r < r_end)
{
if (*l <= *r)
*out++ = *l++;
else
*out++ = *r++;
}
while (l < l_end)
*out++ = *l++;
while (r < r_end)
*out++ = *r++;
if (ENABLE_TRACE)
dump_array("<<-- merge", result, m+n);
}
};
#include <string>
int main()
{
for (size_t i = 1; i <= 10; i++)
{
int array1[] = { 9, 3, 5, 7, 1, 8, 0, 6, 2, 4 };
if (i <= sizeof(array1)/sizeof(array1[0]))
{
cout << "\nMerge array of type int of size " << i << "\n\n";
dump_array("Original", array1, i);
merge_sort(array1, i);
dump_array("PostSort", array1, i);
}
}
for (size_t i = 1; i <= 10; i++)
{
double array2[] = { 9.9, 3.1, 5.2, 7.3, 1.4, 8.5, 0.6, 6.7, 2.8, 4.9 };
if (i <= sizeof(array2)/sizeof(array2[0]))
{
cout << "\nMerge array of type double of size " << i << "\n\n";
dump_array("Original", array2, i);
merge_sort(array2, i);
dump_array("PostSort", array2, i);
}
}
for (size_t i = 1; i <= 10; i++)
{
std::string array3[] = { "nine", "three", "five", "seven", "one", "eight", "zero", "six", "two", "four" };
if (i <= sizeof(array3)/sizeof(array3[0]))
{
cout << "\nMerge array type std::string of size " << i << "\n\n";
dump_array("Original", array3, i);
merge_sort(array3, i);
dump_array("PostSort", array3, i);
}
}
for (size_t i = 1; i <= 10; i++)
{
char array4[] = "jdfhbiagce";
if (i <= sizeof(array4)/sizeof(array4[0]))
{
cout << "\nMerge array type char of size " << i << "\n\n";
dump_array("Original", array4, i);
merge_sort(array4, i);
dump_array("PostSort", array4, i);
}
}
return 0;
}