我想制作虚拟列。在list
id主表中,我需要使用'data'连接虚拟列。
Table: columns
+------+---------+-------------+
| [ID] | ID_USER | DATENAME |
+------+---------+-------------+
| 1 | NULL | Description |
+------+---------+-------------+
| 2 | NULL | Cost |
+------+---------+-------------+
| 3 | 2 | Width |
+------+---------+-------------+
Table: list
+----+-----------+------------+--------+
|[ID]| NAME | DATE | COLOR |
+----+-----------+------------+--------+
| 1 | 234/2016 | 2016-06-06 | red |
+----+-----------+------------+--------+
| 2 | 1000/2016 | 2016-06-07 | blue |
+----+-----------+------------+--------+
| 3 | 3456/2016 | 2016-06-08 | yellow |
+----+-----------+------------+--------+
Table: data
+-----------+-----------+-------+
| ID_REPAIR | ID_COLUMN | VALUE |
+-----------+-----------+-------+
| 1 | 1 | aaaa |
+-----------+-----------+-------+
| 1 | 2 | 10$ |
+-----------+-----------+-------+
| 2 | 1 | bbbb |
+-----------+-----------+-------+
| 2 | 2 | 20$ |
+-----------+-----------+-------+
| 3 | 1 | cccc |
+-----------+-----------+-------+
| 3 | 2 | 30$ |
+-----------+-----------+-------+
Result:
+------+-----------+------------+--------+-------------+------+
| [ID] | NAME | DATE | COLOR | Description | Cost |
+------+-----------+------------+--------+-------------+------+
| 1 | 234/2016 | 2016-06-06 | red | aaaa | 10$ |
+------+-----------+------------+--------+-------------+------+
| 2 | 1000/2016 | 2016-06-07 | blue | bbbb | 20$ |
+------+-----------+------------+--------+-------------+------+
| 3 | 3456/2016 | 2016-06-08 | yellow | cccc | 30$ |
+------+-----------+------------+--------+-------------+------+
在此查询中,我得到名称columns
:
SELECT * FROM `columns` WHERE `id_user` IS NULL
并且在PHP中将ID保存到变量,但是在测试中我希望生成只有第二列的表。在下面的代码中,我想生成带有Cost
列的结果,但始终是插入Description
列:
SQL:
SELECT `list`.`id`, `name`, `date`, `color`, `data`.`value`
FROM `list`
INNER JOIN `data` ON `list`.`id` = `data`.`id_repair` WHERE `repair_data`.`id_column` = 2
我不知道INNER JOIN如何仅WHERE id_column等于2
答案 0 :(得分:2)
SELECT返回的列数无法在执行时动态确定。必须在查询的SELECT
列表中指定列数以及分配给每列的数据类型和名称(或别名)。让我们从那开始吧。
要让SQL语句返回结果集,如图所示,查询必须采用以下格式:
SELECT l.id
, l.name
, l.date
, l.color
, (expr1) AS `Description`
, (expr2) AS `Cost`
FROM list l
...
返回六个列。
对于用于返回expr1
和expr2
列的表达式Description
和Cost
,有几种方法可以展开实体属性值(EAV)模型回到规范关系模型。
最容易理解,但不一定是最佳选择,是在SELECT列表中使用相关子查询。举个例子:
SELECT l.id
, l.name
, l.date
, l.color
, ( SELECT d.value
FROM `data` d
WHERE d.id_repair = l.id
AND d.id_column = 1
ORDER BY d.value
LIMIT 1
) AS `Description`
, ( SELECT c.value
FROM `data` c
WHERE c.id_repair = l.id
AND c.id_column = 2
ORDER BY c.value
LIMIT 1
) AS `Cost`
, ( SELECT w.value
FROM `data` w
WHERE w.id_repair = l.id
AND w.id_column = 3
ORDER BY w.value
LIMIT 1
) AS `Width`
FROM list l
ORDER BY l.id
请注意,SELECT列表中的相关子查询只能返回一行,并返回单个表达式。也就是说,它返回单个值。
作为替代方案,我们可以使用外连接操作和条件聚合。例如:
SELECT l.id
, l.name
, l.date
, l.color
, MAX(IF(d.id_column=1,d.value,NULL)) AS `Description`
, MAX(IF(d.id_column=2,d.value,NULL)) AS `Cost`
, MAX(IF(d.id_column=3,d.value,NULL)) AS `Width`
FROM list l
LEFT
JOIN data d
ON d.id_repair = l.id
GROUP BY l.id, l.name, l.date, l.color
如果我们需要根据存储在数据库中的表中的信息,使用不同数量的列和用于列的表达式来生成返回的结果 dynamic ,那么我们可以先获取来自数据库的信息,然后使用它来帮助我们构造我们需要执行的实际SQL语句以获得最终结果。
或者,正如许多使用实体属性值(EAV)模型的应用程序那样,我们甚至不会尝试将EAV模型强制回到规范的关系模型中。我们只是让应用程序运行从表中检索信息所需的多个查询。基本上与应用程序如何将行插入表中相反。