我有下表页面状态更新。每行都是状态更改:
CREATE TABLE `t1` (
`page_id` int(11) NOT NULL,
`country` varchar(2) COLLATE utf8mb4_unicode_ci NOT NULL,
`status` varchar(45) COLLATE utf8mb4_unicode_ci DEFAULT NULL,
`date` datetime DEFAULT NULL,
PRIMARY KEY (`page_id`,`country`)) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;
INSERT INTO t1 (page_id, country, status, date)
VALUES (1, 'de', 'deactivated', '2018-09-28 14:52:11' ) ,
(1, 'de', 'activated', '2018-09-28 14:54:18' ) ,
(1, 'de', 'deactivated', '2018-09-28 14:60:12' ) ,
(1, 'de', 'moderated', '2018-09-28 14:54:12' ) ,
(2, 'es', 'deactivated', '2018-09-28 14:52:01' ) ,
(2, 'es', 'activated', '2018-09-28 14:52:07' ) ,
(2, 'es', 'deactivated', '2018-09-28 14:52:11' )
我想以这种格式查看表格,以便直到页面具有特定状态时才显示每一行:
CREATE TABLE `t2` (
`page_id` int(11) NOT NULL,
`country` varchar(2) COLLATE utf8mb4_unicode_ci NOT NULL,
`status` varchar(45) COLLATE utf8mb4_unicode_ci DEFAULT NULL,
`date_from` datetime DEFAULT NULL,
`date_to` datetime DEFAULT NULL,
PRIMARY KEY (`page_id`,`country`)) ENGINE=InnoDB DEFAULT CHARSET=utf8mb4 COLLATE=utf8mb4_unicode_ci;
INSERT INTO t2 (page_id, country, status, date_from, date_to)
VALUES (1, 'de', 'deactivated', '2018-09-28 14:52:11', '2018-09-28 14:54:12' ) ,
(1, 'de', 'moderated', '2018-09-28 14:54:12', '2018-09-28 14:54:18' ) ,
(1, 'de', 'activated', '2018-09-28 14:54:18','2018-09-28 14:60:12' ) ,
(1, 'de', 'deactivated', '2018-09-28 14:60:12','2018-01-25 14:60:12' ) ,
(2, 'es', 'deactivated', '2018-09-28 14:52:01','2018-09-28 14:52:07' ) ,
(2, 'es', 'activated', '2018-09-28 14:52:07', '2018-09-28 14:52:11' ) ,
(2, 'es', 'deactivated', '2018-09-28 14:52:11','2018-01-25 14:52:11' ) ;
问题是我们仍在使用MySQL 5.7,还没有升级到具有cte和window函数的mysql8,这很容易解决问题:
SELECT
country,
page_id,
status,
date as date_from,
COALESCE(MIN(date) OVER(PARTITION BY country, page_id ORDER BY date DESC
ROWS BETWEEN 1 PRECEDING AND 1 PRECEDING),
TIMESTAMPADD(day, 1, current_timestamp())) as date_to
FROM t1 ;
我认为应该有一个自我连接的技巧,但无法弄清楚到底是什么!
答案 0 :(得分:1)
在MySQL 8+中,通常为此使用LEAD()
,而不是复杂的表达式。
在早期版本中,您可以使用相关子查询:
select t1.*,
(select tt1.date
from t1 tt1
where tt1.country = t1.country and tt1.page_id = t1.page_id and
tt1.date > t1.date
order by tt1.date asc
limit 1
) as date_to
from t1;
这在每个序列末尾的NULL
值中使用date_to
。这对我来说更有意义。
如果您想重复date_from
(我不建议这样做),则可以执行以下操作:
select t1.*,
(select min(tt1.date_from, t1.date_from)
from t1 tt1
where tt1.page_id = t1.page_id and
tt1.date_from > t1.date_from
) as date_to
from t1;