SQL 使用 N-1 Dense Rank 过滤掉 dense rank N Mysql 中的行

Question

如果我的标题不清楚，这里有一个我正在谈论的例子。

假设我有一个 table 看起来像这样：

+----+------+----+----+---------+-------+---------+-------+
| ID | Time | X  | Y  | X_START | X_END | Y_START | Y_END |
+----+------+----+----+---------+-------+---------+-------+
|  1 |    0 |  8 |  6 |       6 |    10 |       4 |     8 |
|  2 |    0 | 20 | 10 |      18 |    22 |       8 |    12 |
|  3 |    1 |  8 |  8 |       6 |    10 |       6 |    10 |
|  4 |    1 | 10 | 24 |       8 |    12 |      22 |    26 |
+----+------+----+----+---------+-------+---------+-------+

如果我将此查询应用于 table：

WITH 
cte1 AS (SELECT *,
    DENSE_RANK() OVER (ORDER BY TIME) AS DENSE_RANK
    FROM data
    ORDER BY TIME)

SELECT * FROM cte1

我得到：

+----+------+----+----+---------+-------+---------+-------+------------+
| ID | Time | X  | Y  | X_START | X_END | Y_START | Y_END | DENSE_RANK |
+----+------+----+----+---------+-------+---------+-------+------------+
|  1 |    0 |  8 |  6 |       6 |    10 |       4 |     8 |          1 |
|  2 |    0 | 20 | 10 |      18 |    22 |       8 |    12 |          1 |
|  3 |    1 |  8 |  8 |       6 |    10 |       6 |    10 |          2 |
|  4 |    1 | 10 | 24 |       8 |    12 |      22 |    26 |          2 |
+----+------+----+----+---------+-------+---------+-------+------------+

现在我要做的是过滤掉 X 在 X_START - X_END 范围内并且 Y 在 Y_START - Y_END 范围内的任何行AND DENSE_RANK 是 n-1

所以我想要这样的结果：

+----+------+----+----+---------+-------+---------+-------+------------+
| ID | Time | X  | Y  | X_START | X_END | Y_START | Y_END | DENSE_RANK |
+----+------+----+----+---------+-------+---------+-------+------------+
|  1 |    0 |  8 |  6 |       6 |    10 |       4 |     8 |          1 |
|  2 |    0 | 20 | 10 |      18 |    22 |       8 |    12 |          1 |
|  4 |    1 | 10 | 24 |       8 |    12 |      22 |    26 |          2 |
+----+------+----+----+---------+-------+---------+-------+------------+

我是 SQL 的新手，所以我不太确定该怎么做。预先感谢您的所有帮助！

Answer 1

使用 MIN() window 函数确定每个 Time 的最小值 ID 以便在所有其他条件都满足的情况下排除该行：

WITH cte AS (
    SELECT *, 
           DENSE_RANK() OVER (ORDER BY TIME) AS `DENSE_RANK`,
           MIN(ID) OVER (PARTITION BY TIME) min_id
    FROM data
  )
SELECT ID, Time, X, Y, X_START, X_END, Y_START, Y_END, `DENSE_RANK` 
FROM cte
WHERE `DENSE_RANK` = 1 
  OR NOT (ID = min_id AND X BETWEEN X_START AND X_END AND Y BETWEEN Y_START AND Y_END)
ORDER BY `DENSE_RANK`, ID;

参见demo。

SQL 使用 N-1 Dense Rank 过滤掉 dense rank N Mysql 中的行

SQL using N-1 Dense Rank to filter out rows in dense rank N Mysql

mysql

sql

mariadb

window-functions