MYSQL 使用计算字段使用 Group By 或从先前的 MAX 返回值

Question

嘿 stackoverflow 社区，

我目前确实需要 MYSQL 查询的提示。

在我的数据库中，存储了不同“标签”的计数器。

目前我已经获得了当月的最小值和最大值。我的挑战是，我需要上个月的最后一个 MAX。

有人有想法吗？

SELECT 
  CONCAT(YEAR(timestamp),"-",LPAD(MONTH(timestamp), 2, '0')) as DATE,
  MIN(counter) as counterMIN,
  MAX(counter) as counterMAX,
  MIN(timestamp) as timestampMIN,
  MAX(timestamp) as timestampMAX,
  tag
FROM `seller` 
WHERE tag LIKE "SYS%"
GROUP BY CONCAT(YEAR(timestamp),"-",MONTH(timestamp)),`tag`
ORDER BY CONCAT(YEAR(timestamp),"-",LPAD(MONTH(timestamp), 2, '0'))

日期	标签	计数器MIN	计数器MAX
2023-12	SYS001	46	312
2024-01	SYS001	476	807

我的挑战是我需要 counterMAX 值是上个月的值。因此，我想使用子查询，但看起来我无法使用计算字段执行子查询。

日期	标签	计数器MIN	计数器MAX	counterMAX上个月
2023-12	SYS001	46	312	0
2024-01	SYS001	476	807	312

MYSQL数据

CREATE TABLE `seller` (
  `counter` decimal(10,0) NOT NULL,
  `tag` varchar(6) NOT NULL,
  `timestamp` timestamp NOT NULL
);
INSERT INTO `seller` (`counter`, `tag`, `timestamp`) VALUES
(46, 'SYS001', '2023-12-10 08:00:00'),
(199, 'SYS001', '2023-12-13 08:00:00'),
(312, 'SYS001', '2023-12-24 08:00:00'),
(476, 'SYS001', '2024-01-08 08:00:00'),
(493, 'SYS001', '2024-01-12 08:00:00'),
(618, 'SYS001', '2024-01-18 08:00:00'),
(807, 'SYS001', '2024-01-28 08:00:00'),
(858, 'SYS001', '2024-02-06 08:00:00'),
(961, 'SYS001', '2024-02-11 08:00:00'),
(1152, 'SYS001', '2024-02-26 08:00:00'),
(1281, 'SYS001', '2024-03-03 08:00:00'),
(1348, 'SYS001', '2024-03-07 08:00:00'),
(1397, 'SYS001', '2024-03-21 08:00:00'),
(1556, 'SYS001', '2024-04-06 07:00:00'),
(1625, 'SYS001', '2024-04-17 07:00:00'),
(172, 'SYS002', '2023-12-10 08:00:00'),
(334, 'SYS002', '2023-12-21 08:00:00'),
(519, 'SYS002', '2023-12-27 08:00:00'),
(630, 'SYS002', '2024-01-14 08:00:00'),
(790, 'SYS002', '2024-01-17 08:00:00'),
(884, 'SYS002', '2024-01-30 08:00:00'),
(928, 'SYS002', '2024-02-12 08:00:00'),
(938, 'SYS002', '2024-02-15 08:00:00'),
(1103, 'SYS002', '2024-02-23 08:00:00'),
(1289, 'SYS002', '2024-02-25 08:00:00'),
(1361, 'SYS002', '2024-02-27 08:00:00'),
(1553, 'SYS002', '2024-03-07 08:00:00'),
(1597, 'SYS002', '2024-03-15 08:00:00'),
(1604, 'SYS002', '2024-03-22 08:00:00'),
(1762, 'SYS002', '2024-03-29 08:00:00');

提示查询还可以解析上个月的最大值

Answer 1

WITH cte AS (
  SELECT DATE_FORMAT(`timestamp`, '%Y-%m') as `DATE`,
         MIN(counter) as counterMIN,
         MAX(counter) as counterMAX,
         MIN(`timestamp`) as timestampMIN,
         MAX(`timestamp`) as timestampMAX,
         tag
  FROM `seller` 
  WHERE tag LIKE "SYS%"
  GROUP BY `DATE`,`tag`
--  ORDER BY `DATE`
)
SELECT *, LAG(counterMAX) OVER (PARTITION BY tag ORDER BY `DATE`) counterMAXPrevMonth
FROM cte
ORDER BY tag, `DATE`

日期	计数器MIN	计数器MAX	时间戳MIN	时间戳最大	标签	counterMAX上个月
2023-12	46	312	2023-12-10 08:00:00	2023-12-24 08:00:00	SYS001	空
2024-01	476	807	2024-01-08 08:00:00	2024-01-28 08:00:00	SYS001	312
2024-02	858	1152	2024-02-06 08:00:00	2024-02-26 08:00:00	SYS001	807
2024-03	1281	1397	2024-03-03 08:00:00	2024-03-21 08:00:00	SYS001	1152
2024-04	1556	1625	2024-04-06 07:00:00	2024-04-17 07:00:00	SYS001	1397
2023-12	172	519	2023-12-10 08:00:00	2023-12-27 08:00:00	SYS002	空
2024-01	630	884	2024-01-14 08:00:00	2024-01-30 08:00:00	SYS002	519
2024-02	928	1361	2024-02-12 08:00:00	2024-02-27 08:00:00	SYS002	884
2024-03	1553	1762	2024-03-07 08:00:00	2024-03-29 08:00:00	SYS002	1361

小提琴