mysql：使用逗号分隔值

Question

我有一个简单的 mysql 表，用作服务器预订系统。如果用户拥有该服务器的有效预订，则他可以登录到该服务器。示例数据可能如下所示：

+----------+----------------+---------------------+---------------------+
| server   | user           | start               | end                 |
+----------+----------------+---------------------+---------------------+
| server01 | Alice,Bob      | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| server02 | Carlos         | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| server03 | Dan,Erin,Frank | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
+----------+----------------+---------------------+---------------------+

要检查用户Carlos是否预订了server02，我使用以下 SQL 查询：

SELECT 0, id FROM booking WHERE user = 'Carlos'
AND (( start < now() and end > now() ) AND server = 'server02') ;

但如果有多个（逗号分隔）用户，则此查询不起作用。如何查看是否Alice有预订server01？

Answer 1

我会过度考虑您的设计，并会更改为以下内容：

用户表

+--------+----------+
| UserID | Username |
+--------+----------+
|    1   |   PAUL   |
+--------+----------+
|    2   |   Eric   |
+--------+----------+

服务器表

+------------+------------------+
|  ServerID  |    Servername    |
+------------+------------------+
|     1      |      Apache      |
+------------+------------------+
|     2      |      Nginx       |
+------------+------------------+

服务器预订

+-------------+-----------------+---------------+-------------+
|   ServerID  |     UserID      |      Start    |     End     |
+-------------+-----------------+---------------+-------------+
|      1      |        1        |   somedate    |   somedate  |
+-------------+-----------------+---------------+-------------+
|      2      |        1        |   somedate    |   somedate  |
+-------------+-----------------+---------------+-------------+
|      1      |        2        |   somedate    |   somedate  |
+-------------+-----------------+---------------+-------------+

Answer 2

虽然您的booking表非常需要对user列进行标准化（严重、严重、可怕、重要等），但您可以在此期间执行以下操作...

SELECT 0, id FROM booking
WHERE ((start < now() and end > now())
AND server = 'server02') AND
LOCATE(',Alice,',CONCAT(',',user,',')) > 0;

试一试 ！！！

警告：请按照 Johnny000 刚刚建议的方式标准化您的数据！！！为什么？？？

您不会user以 CSV 格式搜索任何体面的查询性能。

如果你 normalize user，你的数据应该是这样的：

+----------+--------+---------------------+---------------------+
| server   | user   | start               | end                 |
+----------+--------+---------------------+---------------------+
| server01 | Alice  | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| server01 | Bob    | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| server02 | Carlos | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| server03 | Dan    | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| server03 | Erin   | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| server03 | Frank  | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
+----------+--------+---------------------+---------------------+

如果您进行归一化的重置，您应该得到以下信息：

用户表

+----+--------+
| id | name   |
+----+--------+
|  1 | Alice  |
|  2 | Bob    |
|  3 | Carlos |
|  4 | Dan    |
|  5 | Erin   |
|  6 | Frank  |
+----+--------+

服务器表

+----+----------+
| id | name     |
+---------------+
|  1 | server01 |
|  2 | server02 |
|  3 | server03 |
+----+----------+

预订表

+----------+--------+---------------------+---------------------+
| serverid | userid | start               | end                 |
+----------+--------+---------------------+---------------------+
| 1        | 1      | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| 1        | 2      | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| 2        | 3      | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| 3        | 4      | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| 3        | 5      | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
| 3        | 6      | 2013-10-28 00:00:00 | 2013-12-31 00:00:00 |
+----------+--------+---------------------+---------------------+

您可以booking像这样在表上创建索引：

ALTER TABLE booking
    ADD INDEX user_server_ndx (userid,serverid),
    ADD INDEX server_user_ndx (serverid,userid),
    ADD INDEX start_end_ndx  (start,end),
    ADD INDEX end_ndx (end)
;

也许您不需要所有这些索引，但您可以决定哪些不需要。