LeetCode498

[jhm9595]

Time Complexity : O(N)

[bigelephant29]

I think there's a better way to structure your code.

if (i % 2 == 0) {
	// moving up
	int r = i >= row ? row - 1 : i;
	int c = i < row ? 0 : i - row + 1;
	while(r < row && r >= 0 && c < col && c >= 0) {
		arr[idx++] = matrix[r][c];
		r--;
		c++;
	}
} else {
	// moving down
	int r = i < col ? 0 : i - col + 1;
	int c = i >= col ? col - 1 : i;
	while(r < row && r >= 0 && c < col && c >= 0) {
		arr[idx++] = matrix[r][c];
		r++;
		c--;
	}
}

[bigelephant29]

Another option:

int r, c, deltaR, deltaC;
if (i % 2 == 0) {
	// moving up
	r = i >= row ? row - 1 : i;
	c = i < row ? 0 : i - row + 1;
	deltaR = -1;
	deltaC = 1;
} else {
	// moving down
	r = i < col ? 0 : i - col + 1;
	c = i >= col ? col - 1 : i;
	deltaR = 1;
	deltaC = -1;
}
while(r < row && r >= 0 && c < col && c >= 0) {
	arr[idx++] = matrix[r][c];
	r += deltaR;
	c += deltaC;
}

I personally like the first implementation better, but you can also consider the second one.

The reason why the first implementation is better is that we can declare "r" and "c" with "final". Using "final" as much as possible is a good practice in Java.

You can read: https://stackoverflow.com/questions/154314/when-should-one-use-final-for-method-parameters-and-local-variables

[bigelephant29]

But still, in the second implementation we remove more duplicate code. In this way it's easier for us to make the code into template and write unit tests on them.

[bigelephant29]

Time complexity is not O(N). It's O(NM). Though at the end you put all the values into a single array, the time complexity is dependent upon the original 2D matrix.