Question

我正在使用MySQL并且有3个这样的表：

Create Table users (
  firstName VARCHAR, 
  lastName VARCHAR, 
  userName VARCHAR, 
  email VARCHAR,
  created DATETIME, etc.

Create Table data_2013 (
  uid VARCHAR,
  d1 INT,
  d2 INT,
  d3 INT, etc

Create Table data_2016 (
  uid VARCHAR,
  d1 INT,
  d2 INT,
  d3 INT, etc

两个数据表中的uid与userName表中的users字段相匹配
每个用户在users表中出现两次（或更多），但始终匹配firstName和lastName。
这些用户的子集（大约100个）在＆＃34; data_xxxx＆＃34;表。
对于2013年数据，userName是一个8个字符的字符串。对于2016年数据，userName是他们当前的电子邮件地址（不一定与2013年使用的电子邮件地址相同）。

我可以让所有拥有2016年数据的用户获得如下查询：

SELECT firstName,lastName,userName 
FROM users 
WHERE created > '2016-01-01' 
AND userName IN(SELECT uid FROM data_2016)`

但我现在想要的是一个查询，它会向我提供具有2013年数据的userName用户列表。但是，正如我所说，userName（或uid）不匹配，但firstName和lastName值应该匹配。

我需要这样的东西，伪代码：

SELECT userName 
FROM users 
WHERE created < '2014-01-01' 
and firstName,lastName IN (
    SELECT firstName,lastName 
    FROM users 
    WHERE created > '2016-01-01' 
    AND userName IN(SELECT uid FROM data_2016))

我确定工会或加入是答案，但我无法弄清楚。

任何提示？

由于

修改

以下是来自users表的一些示例数据：


    +------------------------+-----------+----------+------------------------+---------------------+
    | userName               | firstName | lastName | email                  | created             |
    +------------------------+-----------+----------+------------------------+---------------------+
    | rwhite                 | ROBERT    | WHITE    | xxxxxxxxxx@gmail.com   | 2013-08-05 13:13:23 | 
    | rwhite@company.com     | Robert    | White    | rwhite@company.com     | 2016-10-23 20:26:52 | 
    +------------------------+-----------+----------+------------------------+---------------------+

以上用户的2013年数据示例：


    +--------+---------------------+----+----+----+----+----+
    | uid    | created             | d1 | d2 | d3 | d4 | d5 |
    +--------+---------------------+----+----+----+----+----+
    | rwhite | 2013-08-05 13:24:24 | 38 | 31 |  7 | 22 | 46 |
    +--------+---------------------+----+----+----+----+----+

以上用户的2016年数据示例：


    +--------------------+---------------------+----+----+----+----+----+
    | uid                | created             | d1 | d2 | d3 | d4 | d5 |
    +--------------------+---------------------+----+----+----+----+----+
    | rwhite@company.com | 2016-10-24 12:37:29 | 38 | 48 | 59 | 71 | 17 |
    +--------------------+---------------------+----+----+----+----+----+

EDIT2

我忘记了我有一张第四张表，其中包含某些客户的额外数据：

Create Table users_custA (
  userName VARCHAR,
  id_num VARCHAR,
  etc.
)

此表中同一用户的示例：

+--------------------+-----------+
| userName           | id_num    |
+--------------------+-----------+
| rwhite             | N00123450 | 
| rwhite@company.com | N00123450 | 
+--------------------+-----------+

此id_num保证对于给定的人是唯一的（即，R White是users_custA表中有两个条目的单个人。）

问题仍然存在：如何构建一个查询，生成在两个data_xxxx表中都有数据的userNames列表？

Answer 1

通常，期望名称在时间上是唯一且一致的，这有点不可靠，但如果您确定数据中的情况属实，那么您可以像这样调整查询（假设您有不区分大小写的排序规则）：

SELECT userName 
FROM users As u2013
WHERE created >= '2013-01-01' 
AND created < '2014-01-01'
AND EXISTS (
    SELECT 1 
    FROM users As u2016
    WHERE created >= '2016-01-01' 
    AND created < '2017-01-01'
    AND u2016.FirstName = u2013.FirstName
    AND u2016.LastName = u2013.LastName
    AND EXISTS (SELECT 1 FROM data_2016 WHERE data_2016.uid = u2016.userName));

您会使用WHERE EXISTS而不是WHERE ... IN，因为mysql不支持WHERE (col1, col2) IN ...，它只支持单列or so I understand. < / p>

修改

您可以通过这种方式整合users_custA表格以获得更确定的匹配：

Select * From users_custA Where id_num In ( SELECT id_num FROM ( SELECT DISTINCT id_num FROM users As u JOIN users_custA As a On u.userName = a.userName WHERE created >= '2013-01-01' AND created < '2014-01-01' UNION ALL SELECT DISTINCT id_num FROM users As u JOIN users_custA As a On u.userName = a.userName WHERE created >= '2016-01-01' AND created < '2017-01-01') As union_subquery GROUP BY id_num HAVING COUNT(*) = 2);

msyql比较子查询中两个字段的值

1 个答案: