Given a total of n coins find the total number of full staircase rows that can be built.
Staircase rows are those where i-th row has i coins.
For example, given n = 6, return 3 as you can form staircase like this:
*
* *
* * *
Given n = 9, return 3.
*
* *
* * *
* * *
Note that, the 4th row is invalid as it doesn't have 4 coins.
Approach - 1: Binary Search
To build a staircase till k-th row, we need:
1 + 2 + 3 + ... + k = k*(k + 1) / 2 coins.
So we need to find maximum k such that, k*(k + 1) / 2 <= n.
Since n is an increasing function of k, we can use binary search to solve this problem.
We initialize low and high as 0 and n respectively. In each step, we calculate
the value of coins required using the formula n = k*(k + 1) / 2 where k is the
middle element between low and high. If the required coins are greater than
n the value of high is updated to k - 1 and if its less than n, the value
of low is updated to k + 1. Since we reduce the search space by half at each
iteration, the time complexity is O(logN), where N is the number of coins.
C++ code:
#include <iostream>
using namespace std;
int arrangeCoins(int n) {
long low = 0, high = n;
while (low <= high) {
long k = low + (high - low) / 2;
long cur = k * (k + 1) / 2;
if (cur == n) return (int)k;
if (n < cur) {
high = k - 1;
} else {
low = k + 1;
}
}
return (int)high;
}
int main() {
cout << 6 << " " << arrangeCoins(6) << endl;
cout << 9 << " " << arrangeCoins(9) << endl;
}
Python code:
def arrangeCoins(n):
low = 0
high = n
while low <= high:
k = low + (high - low) // 2
cur = k * (k + 1) // 2
if cur == n: return k
if n < cur:
high = k - 1
else:
low = k + 1
return high
if __name__ == '__main__':
print(6, arrangeCoins(6)) # n = 6, prints 3
print(9, arrangeCoins(9)) # n = 9, prints 3
Time Complexity: O(logN) due to binary search
Space Complexity: O(1)
Approach - 2: Math
We have formulated the equation:
k*(k + 1) / 2 <= n
k^2 + k <= 2*n
k^2 + k - 2*n <=0
We can use Sridharacarya's formula to solve this equation:
k = (-1 + sqrt(1 + 8n))/2
C++ code:
#include <iostream>
#include <cmath>
using namespace std;
int arrangeCoins(int n) {
int(-1 + sqrt(1 + (long)8 * n)) / 2;
}
int main() {
cout << 6 << " " << arrangeCoins(6) << endl;
cout << 9 << " " << arrangeCoins(9) << endl;
}
Python code:
def arrangeCoins(n):
return int((-1 + ((1 + 8 * n) ** 0.5)) / 2)
if __name__ == '__main__':
print(6, arrangeCoins(6)) # n = 6, prints 3
print(9, arrangeCoins(9)) # n = 9, prints 3
Time Complexity: O(1)
Space Complexity: O(1)
Top comments (4)
Nice article 😃
I have one suggestion if you want your C++ solution to be more performant, you can use the keyword
constexprand have your functionarrangeCoinsto be executed at compile Time 😉Thanks! I will look into it more!
This is just a simple high-school algebra problem. People really ask this in programming interviews?
If I were asking someone this, I'd react pretty negatively to the binary-search solution -- any competent engineer should realize there's a simple closed-form solution.
It's an easy question though. Can be asked in phone screen or as a warm up question.