指的是我之前的question我问这个新问题有类似的问题。
我有这样的表records
---- --------- ------
id name points
---- --------- ------
1 aaaa 90
2 bbbb 87
3 cccc 90
4 dddd 80
5 eeee 86
6 ffff 82
7 gggg 87
8 hhhh 85
10 iiii 86
11 iiii 86
12 iiii 86
13 iiii 86
14 iiii 87
15 iiii 73
16 iiii 86
17 iiii 73
18 hhhh 85
19 hhhh 73
20 hhhh 73
21 hhhh 70
22 hhhh 65
23 hhhh 70
24 hhhh 50
形成表格，我想选择具有以下条件的记录
对于第一个最高记录，最高点将位于最前面。 （但只有一个记录）。我有两个最高点90的记录，这里应该显示最近的一个（最大ID）
从第二个记录开始，应以（n * 2）的格式检索字段。我的意思是，对于第二个最大记录，我可以允许4个值下降（2 * 2 = 4），在第二个最大记录-10的范围内。即在表格中，我只选择最近的4条记录（最大id），范围为77到87（87-10 = 77）。现在是第3个最大记录，应该小于或等于76（77-1），即在表中它是73.类似于第3个最大值，我可以允许3 * 2 = 6个记录在63-73之内（73 -10 = 63）..等等..
目前，我正在使用我的另一个问题的回答
SELECT id, name, points FROM (
SELECT
t.*
, @n := IF(@prev_points != points, @n + 1, @n) AS n
, @row := IF(@prev_points != points, 1, @row + 1) AS row
, @prev_points := points
FROM
t
, (SELECT @prev_points := null, @n := 1, @row := 0) var_init_subquery
ORDER BY points DESC, id DESC
) sq
WHERE row <= CASE WHEN n = 1 THEN 1 ELSE n * 2 END
;
但它仅限于同一条记录。是否有可能根据需要在此查询中使用count或任何其他相关函数。希望你帮帮我。
输出应该如下所示，
---- --------- ------
id name points
---- --------- ------
3 cccc 90 -- maximum (only one)
14 iiii 87 |
7 gggg 87 | -- 2nd maximum (allow 2*2 =4 only)
2 bbbb 87 |
16 iiii 86 |
20 hhhh 73 |
19 hhhh 73 | -- 3rd maximum (allow 3*2 =6 only)
17 iiii 73 |
14 iiii 73 |
23 hhhh 70 | and so on for 4th and 5th
21 hhhh 70 |
24 hhhh 50 | -- 4th 4*2 = 8
答案 0 :(得分：1)
/*Sample data*/
CREATE TABLE t
(`id` int, `name` varchar(4), `points` int)
;
INSERT INTO t
(`id`, `name`, `points`)
VALUES
(1, 'aaaa', 90),
(2, 'bbbb', 87),
(3, 'cccc', 90),
(4, 'dddd', 80),
(5, 'eeee', 86),
(6, 'ffff', 82),
(7, 'gggg', 87),
(8, 'hhhh', 85),
(10, 'iiii', 86),
(11, 'iiii', 86),
(12, 'iiii', 86),
(13, 'iiii', 86),
(14, 'iiii', 87),
(15, 'iiii', 73),
(16, 'iiii', 86),
(17, 'iiii', 73),
(18, 'hhhh', 85),
(19, 'hhhh', 73),
(20, 'hhhh', 73),
(21, 'hhhh', 70),
(22, 'hhhh', 65),
(23, 'hhhh', 70),
(24, 'hhhh', 50)
;
/*Query*/
(
SELECT id, name, points, 'maximum (only one)' AS maximum, 'just the max' AS group_range
FROM t
ORDER BY points DESC, id DESC
LIMIT 1
)
UNION ALL
(
SELECT id, name, points, CONCAT(n, ' maximum'), CONCAT('range from ', group_max, ' to ', group_max - 10) FROM (
SELECT
t.*
, @n := IF(points < @group_max - 10, @n + 1, @n) AS n
, @group_max := IF(@n != @prev_n, @group_max - 11, @group_max) AS group_max
, @row := IF(@n != @prev_n, 1, @row + 1)
, IF(@row > @n * 2, 0, 1) AS select_it
, @prev_n := @n
FROM
t
, (SELECT @prev_n := 2, @max := points, @group_max := (SELECT points FROM t WHERE points != (SELECT MAX(points) FROM t) ORDER BY points DESC LIMIT 1), @n := 2, @row := 0 FROM t ORDER BY points DESC LIMIT 1) var_init_subquery
WHERE points != @max
ORDER BY points DESC, id DESC
) sq
WHERE select_it = 1
);
注意，我认为你有点搞砸了你想要的结果。添加了两列来证明：）
/*Result*/
| ID | NAME | POINTS | MAXIMUM | GROUP_RANGE |
|----|------|--------|--------------------|---------------------|
| 3 | cccc | 90 | maximum (only one) | just the max |
| 14 | iiii | 87 | 2 maximum | range from 87 to 77 |
| 7 | gggg | 87 | 2 maximum | range from 87 to 77 |
| 2 | bbbb | 87 | 2 maximum | range from 87 to 77 |
| 16 | iiii | 86 | 2 maximum | range from 87 to 77 |
| 20 | hhhh | 73 | 3 maximum | range from 76 to 66 |
| 19 | hhhh | 73 | 3 maximum | range from 76 to 66 |
| 17 | iiii | 73 | 3 maximum | range from 76 to 66 |
| 15 | iiii | 73 | 3 maximum | range from 76 to 66 |
| 23 | hhhh | 70 | 3 maximum | range from 76 to 66 |
| 21 | hhhh | 70 | 3 maximum | range from 76 to 66 |
| 22 | hhhh | 65 | 4 maximum | range from 65 to 55 |
| 24 | hhhh | 50 | 5 maximum | range from 54 to 44 |
澄清后更新：
(
SELECT id, name, points, 'maximum (only one)' AS maximum, 'just the max' AS group_range
FROM t
ORDER BY points DESC, id DESC
LIMIT 1
)
UNION ALL
(
SELECT id, name, points, CONCAT(n, ' maximum'), CONCAT('range from ', group_max, ' to ', group_max - 10) FROM (
SELECT
t.*
, @n := IF(points < @group_max - 10, @n + 1, @n) AS n
, @group_max := IF(@n != @prev_n, points, @group_max) AS group_max
, @row := IF(@n != @prev_n, 1, @row + 1)
, IF(@row > @n * 2, 0, 1) AS select_it
, @prev_n := @n
FROM
t
, (SELECT @prev_n := 2, @max := points, @group_max := (SELECT points FROM t WHERE points != (SELECT MAX(points) FROM t) ORDER BY points DESC LIMIT 1), @n := 2, @row := 0 FROM t ORDER BY points DESC LIMIT 1) var_init_subquery
WHERE points != @max
ORDER BY points DESC, id DESC
) sq
WHERE select_it = 1
);
| ID | NAME | POINTS | MAXIMUM | GROUP_RANGE |
|----|------|--------|--------------------|---------------------|
| 3 | cccc | 90 | maximum (only one) | just the max |
| 14 | iiii | 87 | 2 maximum | range from 87 to 77 |
| 7 | gggg | 87 | 2 maximum | range from 87 to 77 |
| 2 | bbbb | 87 | 2 maximum | range from 87 to 77 |
| 16 | iiii | 86 | 2 maximum | range from 87 to 77 |
| 20 | hhhh | 73 | 3 maximum | range from 73 to 63 |
| 19 | hhhh | 73 | 3 maximum | range from 73 to 63 |
| 17 | iiii | 73 | 3 maximum | range from 73 to 63 |
| 15 | iiii | 73 | 3 maximum | range from 73 to 63 |
| 23 | hhhh | 70 | 3 maximum | range from 73 to 63 |
| 21 | hhhh | 70 | 3 maximum | range from 73 to 63 |
| 24 | hhhh | 50 | 4 maximum | range from 50 to 40 |