Recursively remove all adjacent duplicates

问题

Find an algorithm to recursively remove all adjacent duplicates in a given string this is the original question.I have thought of an algorithm using stacks.

1.Initialize a stack  and a char_popped variable
2.Push the first char of str into the stack.
3.now iterate through the characters of string.
if the top of stack matches with the character 
{
pop the character from the stack,
char_popped=character
}
else
{
  if(character==char_popped)
         {DONT DO ANYTHING}
  else
         push the character on the stack
         char_popped=null
}

Whatever is remaining on stack is the resultant string.

Auxillary space :O(n) Complexity : O(n)

Is this a correct solution and canyone explain the o(n) solution explained in the post?

回答1:

Is this a correct solution

Yes it is. Let me try to sanitize your idea of using stacks taking into account what @rici mentioned.

You want to traverse through the characters of the string once. This is how your algorithm flows:

 1. Initialize a stack containing the string. (Stack A)
 2. Pop top most element of Stack A and push into Stack B thereby initializing Stack B
 3. Pop from Stack B and pop from Stack A and check if they are equal.
 4. If they aren't push both the elements in Stack B. (The popped element from Stack A is pushed later to preserve order)
 5. Do step 3 till Stack A is empty.

 Stack B should contain the chracters with adjacent duplicates removed.

The above algorithm is clearly O(N) because at every step 3, we are definitely popping an element off Stack A and therefore the size of Stack A goes down by 1 after every step 3.

canyone explain the o(n) solution explained in the post?

Coming to the recursive algorithm that you wanted to get a hang on. This is how it goes:

At any point in time the function removeUtil has two values - one is the character array str from which we have to remove adjacent duplicates and the other is the character last_removed.

Pseudo code with some description -

1. It checks if the first two characters of the str are same, if they are:
   update last_removed and call removeUtil again on the remaining characters 
   (hence the str ++)

2. If first two characters are not same, we do the same exercise as in 1
   only this time we start from str[1], hence removeUtil(str+1, last_removed);

3. Now the returned value from call in 2 is stored in a rem_str because we have 
   characters that didn't match like str[0] in 2 that are orphan and need to be appended
   in the final array just like we pushed elements into the stack in your case.

4. Let's assume the string is 'abbac'-

   In the first call we reach here -> removeUtil("bbac", '\0');
   str[0] -> 'a'

   Now this call -> removeUtil("bbac", '\0'); returns "ac"
   and when the recursion unfolds at this point rem_str -> "ac"
   str[0] -> 'a'

   Hence this :
   if (rem_str[0] && rem_str[0] == str[0])
   {
     *last_removed = str[0];
     return (rem_str+1); // In our case this would return "c" to the previous function call
   }

5. if (rem_str[0] == '\0' && *last_removed == str[0])
     return rem_str;

  Consider the string they mention -> "acbbcddc"
  So at some point we have situation where:
    this call happens-> removeUtil("bbcddc", '\0');
    followed by -> removeUtil("cddc", 'b');
    followed by -> removeUtil("ddc", 'b');
    followed by -> removeUtil("c", 'd');
    That returns 'c' and considering the str[0] in the string "cddc" , it goes to case 4
    and therefore 'c' gets removed returning an empty string and last_removed -> 'c'.

  So removeUtil("bbcddc", '\0'); returns an empty string with last_removed -> 'c'.

  However, here str[0] is the same as the last_removed character and therefore should be removed.


6. If none of the following string matches that means str[0] should be appended to final answer.
   rem_str--;
   rem_str[0] = str[0];
   return rem_str;

And here also since we are traversing the character array and 'forming' the final character array that get returned by appropriately appending and removing characters, the time complexity is O(N).

Sidenote- Whenever thinking about recursion, its always easier to think of it like a depth of function calls unfolding and then returning values which then get collected to return the final result.

Anyway, I hope I could clarify myself and hope this gets you started in the right direction.

来源：https://stackoverflow.com/questions/26825084/recursively-remove-all-adjacent-duplicates