Question

As I tried to solve this interview problem: find the sum of continuous element in an array which equals to a target number, so I come up with the following code. However, I really don’t understand why it has some memory allocation problem. Here is the link to the full code. when I tried to insert the second element of currSum into the set and map, it will have some error message like “memory clobbered past end of allocated block”. Isn’t set, map dynamically allocated for insert? I really don’t understand why it’s not working as I think.

I also paste the full code here:

        #include <map>
        #include <set>
        #include <iostream>
        using namespace std;

        void print_array(int arr[], int start, int end)
        {
            for(int i=start;i<=end;i++)
                cout<<arr[i]<<" ";
            cout<<endl;
        }

        //given an array, find the continous sum of the array which is a target number
        void findTargetSum(int arr[], int target, int sizeArr)
        {
            map<int, int> valIdxMap;
            set<int> prevSet;
            int* currSum= new int(sizeArr+1);
            currSum[0]=0;   
            for(int i=1;i<=sizeArr;i++)
            {
                currSum[i]=currSum[i-1]+arr[i-1];
            }

            //now the problem is to find two elements i, j in array currSum,where currSum[j]-currSum[i]=target && j>i
            for(int i=0; i<=sizeArr;i++)
            {
                int tmp=currSum[i]-target;
                set<int>::iterator iter=prevSet.find(tmp);
                if (iter !=prevSet.end())
                {
                    cout<<"located one range of array elements with sum to"<<target<<endl;
                    int startIdx=valIdxMap[*iter];
                    print_array(arr,startIdx,i-1);
                }
                else
                {
                    prevSet.insert(currSum[i]);
                    valIdxMap.insert(make_pair(currSum[i],i));
                }
            }
            delete currSum;
        }

        void testfindTargetSum()
        {
            int tst_arr[]={2,4,5,-1,3,8};
            int target=11;
            findTargetSum(tst_arr,11,6);
        }

        int main()
        {
           testfindTargetSum();
           return 1;
        }

Answer 1

The error is in this line:

int* currSum= new int(sizeArr+1);

That line acquires a single int and initializes it to the value sizeArr+1. You probably meant:

int* currSum= new int[sizeArr+1];

That will acquire a block of sizeArr+1 elements of type int. Additionally you will have to change the line delete currSum; to be delete [] currSum;.

My advice, on the other hand, would be not to manage memory manually but use a standard container, for example:

std::vector<int> currSum( sizeArr+1 );

Will basically be an in-place replacement for your current implementation and it will manage memory automatically.

As of the implementation, I believe that you can actually do it without any extra memory in O(N) by accumulating the start index, end index and sum of the values in three variables while iterating. While the accumulated value is smaller than the target, increment the end index and add the value. When the accumulated value grows higher than the target, increment the start index and decrement the accumulated value by that amount.