Question

I was asked in an interview to find if the single linked list has a loop. I have searched this SO and found few answers Interview question: How to detect a loop in a linked list? and I knew about those answers before but somehow I answered it different which I had in mind from my discussion with friends long time back, but I wasn’t sure if the answer is correct. If try it myself it works as expected. Somehow the interviewer is not happy. Can someone clarify what is wrong in this ?

struct node {
int flag;
struct node *next;
}

My idea here is not to use two pointers and hop variantly but to use single pointer and traverse each node e.g:

node =  node->next;

Only this I would do is set the flag variable to 1 when ever I traverse the node and check each time e.g

if (node->flag)
{ 
  return TRUE; // implies list is a loop
}
else{
   node->flag = 1;
}

Other than efficiency, what is wrong in this approach ?

Answer 1

You are using an extra integer variable flag and say if the sizeof(int) in your machine is 4 bytes and say you have 100,000 nodes in the linked-list, then you will end up using 400,000 bytes more. That will be approximately 390M/.38G

To reset the flag for all the nodes in the linked-list, even though it is O(n), you’ll have to traverse all the 100,000 nodes. Its an overhead.

Now, if you compare your flag based solution with the 2-variable solution, then it is almost the same computational time but 2-variable solution is considered very memory efficient as it does not use 390M of memory to identify the loop!