LeetCode 37. 解数独（dfs + 妙用flag剪枝）

37. 解数独

编写一个程序，通过填充空格来解决数独问题。

数独的解法需 遵循如下规则：

数字 1-9 在每一行只能出现一次。
数字 1-9 在每一列只能出现一次。
数字 1-9 在每一个以粗实线分隔的 3x3 宫内只能出现一次。（请参考示例图）
数独部分空格内已填入了数字，空白格用 '.' 表示。

示例1：

输入：
board = [["5","3",".",".","7",".",".",".","."],
["6",".",".","1","9","5",".",".","."],
[".","9","8",".",".",".",".","6","."],
["8",".",".",".","6",".",".",".","3"],
["4",".",".","8",".","3",".",".","1"],
["7",".",".",".","2",".",".",".","6"],
[".","6",".",".",".",".","2","8","."],
[".",".",".","4","1","9",".",".","5"],
[".",".",".",".","8",".",".","7","9"]]

输出：
[["5","3","4","6","7","8","9","1","2"],
["6","7","2","1","9","5","3","4","8"],
["1","9","8","3","4","2","5","6","7"],
["8","5","9","7","6","1","4","2","3"],
["4","2","6","8","5","3","7","9","1"],
["7","1","3","9","2","4","8","5","6"],
["9","6","1","5","3","7","2","8","4"],
["2","8","7","4","1","9","6","3","5"],
["3","4","5","2","8","6","1","7","9"]]

解释： 输入的数独如上图所示，唯一有效的解决方案如下所示：

---

沿用上一题的思路👉LeetCode 36. 有效的数独。

依旧使用三个数组分别表示图中某点(x,y)在行、列、宫内是否出现过。

显然题目的答案是一个关于填入数独图中的空白的组合决策，所以单独研究这些空白的点，并将其存入到blank的List中，然后考虑使用深搜，每个空都可枚举一遍1~9之间的数。

如果满足要求（数独的规则）就接着移到下一个空白处(dfs(deep + 1))，直到dfs的深度与空白的点的个数相同，表明此时已到达终点，可以退出。

问题是数独的解是唯一的，即dfs有一个分支到达终点得解后board就不允许再被改动了，所以其他分支就得停止搜索。

为了解决这个问题，引入一个终点标记flag，初始的值为false，然后在每次枚举1~9的时候检验一遍是否满足!flag，当深搜到终点时置其值为true，这样就可以使其他分支执行for循环时自动退出return（剪枝）。

---

Java代码

class pii{      //空白的点
    int first;
    int second;
    public pii(int first_, int second_){
        first = first_;
        second = second_;
    }
}

class Solution {
    private List<pii> blank;
    private boolean[][] row;
    private boolean[][] col;
    private boolean[][][] g;
    private boolean flag;

    public void init(char[][] board){
        for(int i = 0;i < 9;i ++)
            for(int j = 0;j < 9;j ++){
                char c = board[i][j];
                if(c == '.')
                    blank.add(new pii(i, j));
                else{
                    row[i][c - '0'] = true;
                    col[j][c - '0'] = true;
                    g[i / 3][j / 3][c - '0'] = true;
                }
            }
    }

    public void dfs(int deep, char[][] board){
        if(deep == blank.size()){
            flag = true;       //当找到了解，则flag标记可以终止掉其他dfs()函数执行时的for循环(不断回溯)
            return;
        }
        pii node = blank.get(deep);
        for(int i = 1;i <= 9 && !flag;i ++){
            int x = node.first, y = node.second;
            if(!row[x][i] && !col[y][i] && !g[x / 3][y / 3][i]){
                row[x][i] = col[y][i] = g[x / 3][y / 3][i] = true;      
                board[x][y] = (char) (i + '0');        //更改状态

                dfs(deep + 1, board);

                row[x][i] = col[y][i] = g[x / 3][y / 3][i] = false;     //撤销状态
            }
        }
    }

    public void solveSudoku(char[][] board) {
        blank = new ArrayList<>();
        row = new boolean[10][10];
        col = new boolean[10][10];
        g = new boolean[3][3][10];
        flag = false;
        init(board);

        dfs(0, board);
    }
}