Mysql 连接表-有点复杂

每周，我的客户都会通过第三方营销团队进行广播和电子邮件宣传，以推广他的产品。他想知道自己每周挣多少钱/花多少钱，并将数据细分。表数据如下所示：

tablename: orders
+----------+--------------+---------------------+------------+
| order_id | order_amount | referral            | order_date |
+----------+--------------+---------------------+------------+
| 1        | 16.00        | Radio: Jack Johnson | 2013-07-01 |
| 2        | 50.00        | Radio: Jack Johnson | 2013-07-02 |
| 3        | 16.00        | Radio: Jack Johnson | 2013-07-02 |
| 4        | 50.00        | Radio: Thabo Mbeki  | 2013-07-03 |
| 5        | 16.00        | Email: Renewal      | 2013-07-08 |
| 6        | 50.00        | Other               | 2013-07-08 |
| 7        | 16.00        | Email: Renewal      | 2013-07-08 |
| 8        | 50.00        | Radio: Thabo Mbeki  | 2013-07-09 |
+----------+--------------+--------------+-------------------+

tablename: media
+----------+-------+-------+--------------+-------+------------+
| media_id | spots | cost  | name         | type  | date       |
+----------+-------+-------+--------------+-------+------------+
| 1        | 10    | 10000 | Jack Johnson | radio | 2013-07-01 |
| 2        | 5     | 5000  | Thabo Mbeki  | radio | 2013-07-01 |
| 3        | 10    | 10000 | Jack Johnson | radio | 2013-07-08 |
| 4        | 5     | 5000  | Thabo Mbeki  | radio | 2013-07-08 |
| 5        | 1     | 250   | Renewal      | email | 2013-07-08 |
+----------+---------------+--------------+--------------------+

+------------+--------------+-------+------+----------+---------+--------------+----------------+
| week       | name         | type  |spots | cost     | revenue | total orders | avg. per order |
+------------+--------------+-------+------+----------+---------+--------------+----------------+
| 2013-07-01 | Jack Johnson | radio | 10   | 10000    | 82.00   | 3            | 27.33          |
| 2013-07-01 | Thabo Mbeki  | radio | 5    | 5000     | 50.00   | 1            | 50.00          |
+------------+--------------+-------+------+----------+---------+--------------+----------------+

我需要生成如下所示的HTML表：

tablename: orders
+----------+--------------+---------------------+------------+
| order_id | order_amount | referral            | order_date |
+----------+--------------+---------------------+------------+
| 1        | 16.00        | Radio: Jack Johnson | 2013-07-01 |
| 2        | 50.00        | Radio: Jack Johnson | 2013-07-02 |
| 3        | 16.00        | Radio: Jack Johnson | 2013-07-02 |
| 4        | 50.00        | Radio: Thabo Mbeki  | 2013-07-03 |
| 5        | 16.00        | Email: Renewal      | 2013-07-08 |
| 6        | 50.00        | Other               | 2013-07-08 |
| 7        | 16.00        | Email: Renewal      | 2013-07-08 |
| 8        | 50.00        | Radio: Thabo Mbeki  | 2013-07-09 |
+----------+--------------+--------------+-------------------+

tablename: media
+----------+-------+-------+--------------+-------+------------+
| media_id | spots | cost  | name         | type  | date       |
+----------+-------+-------+--------------+-------+------------+
| 1        | 10    | 10000 | Jack Johnson | radio | 2013-07-01 |
| 2        | 5     | 5000  | Thabo Mbeki  | radio | 2013-07-01 |
| 3        | 10    | 10000 | Jack Johnson | radio | 2013-07-08 |
| 4        | 5     | 5000  | Thabo Mbeki  | radio | 2013-07-08 |
| 5        | 1     | 250   | Renewal      | email | 2013-07-08 |
+----------+---------------+--------------+--------------------+

+------------+--------------+-------+------+----------+---------+--------------+----------------+
| week       | name         | type  |spots | cost     | revenue | total orders | avg. per order |
+------------+--------------+-------+------+----------+---------+--------------+----------------+
| 2013-07-01 | Jack Johnson | radio | 10   | 10000    | 82.00   | 3            | 27.33          |
| 2013-07-01 | Thabo Mbeki  | radio | 5    | 5000     | 50.00   | 1            | 50.00          |
+------------+--------------+-------+------+----------+---------+--------------+----------------+

---

对于我来说，我试图弄清楚如何连接表，并根据订单中的引用字段对它们进行分组，这变得很复杂。由于orders.reference与media.name不匹配，我猜有一个WHERE引用，如%name，但我不确定如何执行该操作。

如果忽略其他，则最好使用join WHERE子句

WHERE orders.order_date = media.date
 AND  LOWER(orders.referral)   = LOWER(media.type + ": " + media.name)

这是Sybase语法，如果您需要对MySQL进行一些调整，我很抱歉

要处理其他问题，只需对媒体表进行外部连接

类似“未检查正确性”的内容将按日期生成报告；如果你愿意的话，你需要每周对小组进行进一步的按摩。这并不难，但我觉得这超出了你问题的主要范围

SELECT SUM(order.amount) AS revenue, COUNT(order.amount) as total_orders
      ,SUM(order.amount)/COUNT(order.amount) as average_order
      ,SUM(spots) spots, media.type, media.name, order_date
FROM   order
LEFT OUTER JOIN media 
ON        orders.order_date = media.date
     AND  LOWER(orders.referral)   = LOWER(media.type + ": " + media.name)
GROUP BY order.referral, order.date
-- May be add to group by:    , COALESCE(media.type,media.name,"Other")

---

我认为这应该是可行的：

SELECT m.name, [other needed columns]
FROM media m JOIN orders o ON o.referral LIKE '%' + m.name +'%'
WHERE m.date between [date] and [date]
GROUP BY m.name, [other groupings]

---

希望这有帮助

这里帮了大忙。这是我最后使用的查询：

SELECT m.date, m.name, m.type, m.spots, m.cost, 
       SUM(o.order_amount) as revenue, 
       COUNT(o.order_amount) as orders,
       SUM(o.order_amount)/COUNT(o.order_amount) as average_order 
FROM   media m
JOIN   orders o
ON     LOWER(o.referral) = CONCAT(LOWER(m.type), ': ' ,LOWER(m.name))
WHERE  (DATE(o.order_date) BETWEEN '2013-06-24' AND '2013-06-30')
GROUP BY m.name, m.type

---

您是否可以选择修复表以将引用分为两列？是的，数据需要清理。他们有这么多的命名约定，我永远不会给他们想要看到的准确报告。我正在为所有表实现一个适当的命名方案，这将在查询和结果中产生巨大的差异。我的意思是，orders表中的列和media表中的列具有相同的单独名称和类型。加入将更容易是的，我理解，这将使事情变得更容易。使用下面的解决方案，我能够使用CONCAT将“类型”和“名称”组合起来，以重新创建“引用”。但是在orders表中将它们分成两个字段就不需要这样做了。这非常有帮助！谢谢你把它放在一起。谢谢你把金额加起来，然后再计算我无法直接思考的金额。这是一个伟大的步骤。我现在也知道，如果他们想要好的匹配，数据需要清理。