Question

我有一个包含数百万行的表，结构如下：

UserID | Date       | Points
1      | 2016-05-01 | 240
1      | 2016-05-02 | 500
1      | 2016-05-03 | 650
2      | 2016-05-01 | 122
2      | 2016-05-02 | 159
2      | 2016-05-03 | 290

依此类推。

我需要在Points和2016-05-03之间找到2016-05-01对每个ID的差异，按Points中的差异排序，然后返回前100个。我的数据库包含数亿行，因此需要快速操作。

我从哪里开始？我在看group_concat，但我不确定这个用例是否正确。

Answer 1

此查询正在运行，速度不是很快，但可以制作。

select tbl.id as UserId, ((select t.points from tbl t where date = '2016-05-03' and t.id=tbl.id)-(select t.points from tbl t where date = '2016-05-01' and t.id=tbl.id)) as diff from tbl group by id order by diff limit 100;

运行此操作并告诉我这是否需要花费很多时间。

Answer 2

你应该有如下的子查询

SELECT 
     user_id,
     (select points from tbl_test where tbl_test.date_of='2016-05-03' and user_id=t1.user_id)-t1.points as difference_in_points 
FROM `tbl_test` as t1 
WHERE date_of='2016-05-01' order by difference_in_points desc 
limit 100

Answer 3

自己加入桌子。

SELECT t1.user_id, t1.points - t2.points as diff
FROM yourTable AS t1
JOIN yourTable AS t2 ON t1.user_id = t2.user_id
WHERE t1.date = '2016-05-03'
AND t2.date = '2016-05-01'
ORDER BY diff DESC
LIMIT 100

在数百万的数据库中组合多行

3 个答案: