LeetCode 231 - Power of Two

https://leetcode.com/problems/power-of-two/

Given an integer, write a function to determine if it is a power of two.

Example 1:

Input: 1
Output: true 
Explanation: 20 = 1

Example 2:

Input: 16
Output: true
Explanation: 24 = 16

Example 3:

Input: 218
Output: false

https://leetcode.com/problems/power-of-two/discuss/63966/4-different-ways-to-solve-Iterative-Recursive-Bit-operation-Math

Method 1: Iterative

check if n can be divided by 2. If yes, divide n by 2 and check it repeatedly.

if (n == 0) return false;
while (n%2 == 0) n/=2;
return n == 1;

Time complexity = O(log n)

Method 2: Recursive

return n > 0 && (n == 1 || (n%2 == 0 && isPowerOfTwo(n/2)));

Time complexity = O(log n)

Method 3: Bit operation

If n is the power of two:

• n = 2 ^ 0 = 1 = 0b0000...00000001, and (n - 1) = 0 = 0b0000...0000.
• n = 2 ^ 1 = 2 = 0b0000...00000010, and (n - 1) = 1 = 0b0000...0001.
• n = 2 ^ 2 = 4 = 0b0000...00000100, and (n - 1) = 3 = 0b0000...0011.
• n = 2 ^ 3 = 8 = 0b0000...00001000, and (n - 1) = 7 = 0b0000...0111.

we have n & (n-1) == 0b0000...0000 == 0

Otherwise, n & (n-1) != 0.

For example, n =14 = 0b0000...1110, and (n - 1) = 13 = 0b0000...1101.

return n > 0 && ((n & (n-1)) == 0);

Time complexity = O(1)

Method 4: Math derivation

Because the range of an integer = -2147483648 (-2^31) ~ 2147483647 (2^31-1), the max possible power of two = 2^30 = 1073741824.

(1) If n is the power of two, let n = 2^k, where k is an integer.

We have 2^30 = (2^k) * 2^(30-k), which means (2^30 % 2^k) == 0.

(2) If n is not the power of two, let n = j*(2^k), where k is an integer and j is an odd number.

We have (2^30 % j*(2^k)) == (2^(30-k) % j) != 0.

return n > 0 && (1073741824 % n == 0);

Time complexity = O(1)

--
Update:
Thanks for everyone's comment. Following are some other solutions metioned in comments.

Method 5: Bit count

Very intuitive. If n is the power of 2, the bit count of n is 1.
Note that 0b1000...000 is -2147483648, which is not the power of two, but the bit count is 1.

return n > 0 && Integer.bitCount(n) == 1;

Time complexity = O(1)
The time complexity of bitCount() can be done by a fixed number of operations.
More info in https://stackoverflow.com/questions/109023.

Method 6: Look-up table
There are only 31 numbers in total for an 32-bit integer.

return new HashSet<>(Arrays.asList(1, 2, 4, 8, 16, 32, 64, 128, 256, 512, 1024, 2048, 4096, 8192, 16384, 32768, 65536, 131072, 262144, 524288, 1048576, 2097152, 4194304, 8388608,16777216, 33554432, 67108864, 134217728, 268435456, 536870912, 1073741824)).contains(n);

time complexity = O(1)

https://leetcode.com/problems/power-of-two/discuss/63974/Using-nand(n-1)-trick

Power of 2 means only one bit of n is '1', so use the trick n&(n-1)==0 to judge whether that is the case

    bool isPowerOfTwo(int n) {
        if(n<=0) return false;
        return !(n&(n-1));
    }

Another trick using n & -n which get the right most 1 bit (set bit) of n, for example:
n = 5, -n = -5
n & -n = 0101 & 1011 = 0001, the right most set bit of 5 is 0001.
Another example:
n = 12, -n = -12
n & -n = 1100 & 0100=0100, as you can see the right most set bit of 12 is 0100.

    public boolean isPowerOfTwo(int n) {
        return n > 0 && (n & -n) == n;
    }

https://leetcode.com/problems/power-of-two/discuss/64127/One-line-Java-solution

 public boolean isPowerOfTwo(int n) {
    return ((n & (n-1))==0 && n>0);
}