Question

我有一张这样的表：

// requests
+----+----------------+----------------+-------------+
| id | user_id        |      ip        |  unix_time  |
+----+----------------+----------------+-------------+
| 1  | 12353          | NULL           | 1339412843  |
| 2  | 12353          | NULL           | 1339412864  |
| 3  | NULL           | 178.253.29.175 | 1339412894  |
| 4  | 3422           | NULL           | 1339412899  |
| 5  | 3422           | NULL           | 1339412906  |
| 6  | 3422           | NULL           | 1339412906  |
| 7  | NULL           | 148.23.29.109  | 1339413640  |
| 8  | NULL           | 148.23.29.109  | 1339413621  |
| 9  | 5462           | NULL           | 1339414490  |
| 10 | NULL           | 178.253.29.175 | 1339419901  |
| 11 | 8007           | NULL           | 1339424860  |
| 12 | 8007           | NULL           | 1339424822  |
| 13 | 12353          | NULL           | 1339424902  |
| .  | .              | .              | .           |
| .  | .              | .              | .           |
| .  | .              | .              | .           |
+----+----------------+----------------+-------------+

我也有这张桌子：

// per_days
+----+---------+--------------+----------------+-----------------+--------------+
| id | user_id |  AllVisited  | MaxConsecutive | LastConsecutive | request_numb |
+----+---------+--------------+----------------+-----------------+--------------+
| 1  | 12353   | 43           | 8              | 3               | 47           |
| 2  | 3422    | 530          | 130            | 32              | 100          |
| .  | .       | .            | .              | .               | .            |
| .  | .       | .            | .              | .               | .            |
| .  | .       | .            | .              | .               | .            |
+----+---------+--------------+----------------+-----------------+--------------+
-- each user has one row into this ^ table (I mean user_id column is unique)

我需要一个查询来选择requests表中最后一天的所有行（我将每天通过事件执行该查询），然后更新{的所有行{1}}表（分别为每个用户）。像这样：

per_days

如何修复查询？

Answer 1

使用子查询加入，该子查询返回时间范围内requests中的所有用户ID。使用LEFT JOIN，然后您的if there's a row测试变为r.user_id IS NOT NULL。

UPDATE 
     per_days AS p
LEFT JOIN (
    SELECT DISTINCT user_id
    FROM requests 
    WHERE unix_time > subdate(now(), interval '1' day)) AS r
ON   p.user_id = r.user_id
SET  p.AllVisited      = p.AllVisited + IF(r.user_id IS NOT NULL, 1, 0),
     p.MaxConsecutive  = IF( p.LastConsecutive > p.MaxConsecutive, LastConsecutive, MaxConsecutive),    
     p.LastConsecutive = IF(r.user_id IS NOT NULL, p.LastConsecutive + 1, 1)

如何根据另一个表的值更新表的所有行？

1 个答案: