from heapq import heappush, heappop

def solve(N, S):
    # Convert string to array: '0' -> +1, '1' -> -1
    A = [1 if c == '0' else -1 for c in S]
    
    # Priority queue to store indices of '0's (max heap using negative indices)
    pq = []
    
    # Running prefix sum
    prefix_sum = 0
    
    # List to store prefix sums up to current position
    prefix_sums = [0]  # Initial prefix sum before first character
    
    # Count of operations
    operations = 0
    
    # Process each character
    for i in range(N):
        # Update prefix sum with current character's value
        prefix_sum += A[i]
        
        # If current character is '0', add its index to priority queue
        if A[i] == 1:
            heappush(pq, -i)  # Negative index for max heap behavior
        
        # Check if current prefix sum violates the "good" condition
        min_prev = min(prefix_sums)
        while prefix_sum > min_prev and pq:
            # Get the rightmost '0' index
            idx = -heappop(pq)
            # Flip '0' to '1': +1 becomes -1, reducing prefix sum by 2
            A[idx] = -1
            prefix_sum -= 2
            operations += 1
        
        # Store the current prefix sum
        prefix_sums.append(prefix_sum)
        # Optional: Keep prefix_sums manageable by removing oldest if unnecessary
        if len(prefix_sums) > i - 1:
            prefix_sums.pop(0)
    
    return operations

# Input handling
N = int(input())
S = input().strip()

# Output the result
print(solve(N, S))