Find maximum in stack in O(1) without using additional stack in Python

The task is to design a stack which can get the maximum value in the stack in O(1) time without using an additional stack in Python.

Examples:

Input: Consider the following SpecialStack

16  –> TOP
29
15
19
18
When getMax() is called it should return 29, 
which is the maximum element in the current stack. 

If we do pop two times on stack, the stack becomes

15  –> TOP
19
18

When getMax() is called, it should return 19 
which is the maximum in the current stack.

Approach: Instead of pushing a single element to the stack, push a pair instead. The pair consists of the (value, localMax) where localMax is the maximum value upto that element. 

  • When we insert a new element, if the new element is greater than the local maximum below it, we set the local maximum of a new element equal to the element itself.
  • Else, we set the local maximum of the new element equal to the local maximum of the element below it.
  • The local maximum of the top of the stack will be the overall maximum.
  • Now if we want to know the maximum at any given point, we ask the top of the stack for local maximum associated with it which can be done in O(1).

Below is the implementation in Python:

Python
# Python3 implementation of the approach
class Block:

    # A block has two elements
    # as components (i.e. value and localMax)
    def __init__(self, value, localMax):
        self.value = value
        self.localMax = localMax


class Stack:
    def __init__(self, size):

        # Setting size of stack and
        # initial value of top
        self.stack = [None] * size
        self.size = size
        self.top = -1

    # Function to push an element
    # to the stack
    def push(self, value):

        # Don't allow pushing elements
        # if stack is full
        if self.top == self.size - 1:
            print("Stack is full")
        else:
            self.top += 1

            # If the inserted element is the first element
            # then it is the maximum element, since no other
            # elements is in the stack, so the localMax
            # of the first element is the element itself
            if self.top == 0:
                self.stack[self.top] = Block(value, value)

            else:

                # If the newly pushed element is less
                # than the localMax of element below it,
                # Then the over all maximum doesn't change
                # and hence, the localMax of the newly inserted
                # element is same as element below it
                if self.stack[self.top - 1].localMax > value:
                    self.stack[self.top] = Block(
                        value, self.stack[self.top - 1].localMax)

                # Newly inserted element is greater than
                # the localMax below it, hence the localMax
                # of new element is the element itself
                else:
                    self.stack
                    self.stack[self.top] = Block(value, value)

            print(value, "inserted in the stack")

    # Function to remove an element
    # from the top of the stack
    def pop(self):

        # If stack is empty
        if self.top == -1:
            print("Stack is empty")

        # Remove the element if the stack
        # is not empty
        else:
            self.top -= 1
            print("Element popped")

    # Function to find the maximum
    # element from the stack
    def getMax(self):

        # If stack is empty
        if self.top == -1:
            print("Stack is empty")
        else:

            # The overall maximum is the local maximum
            # of the top element
            print("Maximum value in the stack:",
                  self.stack[self.top].localMax)

# Driver code


# Create stack of size 5
stack = Stack(5)
stack.push(2)
stack.getMax()
stack.push(6)
stack.getMax()
stack.pop()
stack.getMax()

Output
2 inserted in the stack
Maximum value in the stack: 2
6 inserted in the stack
Maximum value in the stack: 6
Element popped
Maximum value in the stack: 2

Time Complexity: O(1)
Auxiliary Space: O(N)