Leetcode 1536. Minimum Swaps to Arrange a Binary Grid (python)

题目

解法：

将2D问题转化为1D，然后贪心的移动

class Solution:
    def minSwaps(self, grid: List[List[int]]) -> int:
        n = len(grid)
        # take this as a 1-d problem,count the continues 0 from the end in every row. What we need to do is to put the rows at the correct posistion. 
        zero_count = []
        for i in range(n):
            count = 0
            for j in range(n-1,-1,-1):
                if grid[i][j]:
                    break
                count += 1
            zero_count.append(count)
        
        # for ith from the first row to the last, we greedyly find the satisfied row to put at the ith position. When i is at the end, we have placed all the rows at correct position
        # moves counts the total swaps needed
        moves = 0
        for i in range(n):
            if zero_count[i]<n-i-1:
                # i is the current row position we need to put the satisfied row
                pos = i
                # we move down until find the satisfied row
                while pos<n and zero_count[pos]<n-i-1:
                    pos += 1
                # if we can't find the satisfied row for current i, return -1
                if pos==n:
                    return -1
                # we swap the pos row with the pos-1 row until to the ith row. But here we actually only need to swap the zero_count list
                while pos>i:
                    zero_count[pos],zero_count[pos-1] = zero_count[pos-1],zero_count[pos]
                    pos -= 1
                    moves+=1
        return moves

二刷
两个问题需要证明：

1. 为什么往后找到第一个符合条件的行就可以，而不是需要找到刚好符合的那一行。比如当前行需要10个tailing zero，我们最先找到的一行有1000个tailing zero，那么把这行作为目标行是肯定没有问题的，因为当前行之后的所有行需要的tailing zero数量肯定比当前行少，所以可以优先满足当前行，这是greedy的思想（
   Suppose, current row (i-th row) asks for a row with at least 4 zeros, and the neaerst downward row (j-th) has 5 tailing zeros. You can greedily but aslo safely choose it as candidate. any rows after than i-th row will ask for less than 4 tailing rows. So you greedy approach will cause no trouble at all.
   ）
2. 为什么这么移动是最优的，感觉不太好证明，因为每次移动都会改变其它行的位置，所以一开始才会想到用最短路径的方法来解决

class Solution {
public:
    int minSwaps(vector<vector<int>>& grid) {
        int n = grid.size();
        vector<int> zero_count(n,0);
        for(int i=0;i<n;i++){
            int cnt = 0;
            for(int j=n-1;j>=0;j--){
                if(grid[i][j] != 0){
                    break;
                }
                cnt++;
            }
            zero_count[i] = cnt;
            // cout << cnt << endl;
        }
        int moves = 0;
        for(int i=0;i<n;i++){
            int curr_row = i;
            while(curr_row < n && zero_count[curr_row] < n-i-1) curr_row++;
            if(curr_row == n) return -1;
            while(curr_row > i){
                swap(zero_count[curr_row],zero_count[curr_row-1]);
                curr_row--;
                moves++;
            }
        }
        return moves;
    }
};

时间复杂度:O(n^2), 来自于第二个for循环，每行访问一次，每次访问至多移动n个位置
空间复杂度:O(n)