我有下表

 id  |   add_card_id |   rate | category |  category_count |    counts | total
7597 |    723        |    3   |   7      |      1          |     300   | 108
7596 |    723        |    3   |   9      |      1          |     500   | 180
7594 |    723        |    3   |   11     |      1          |     400   | 240
7593 |    723        |    5   |   10     |      2          |     400   | 240
7593 |    723        |    5   |   13     |      2          |     400   | 240


我想获取total(列名)的总和,其中rate相同,但总和仅包括第n个。在具有最大总数的行中,第n个由category_count(列名)定义。

例如,前三列的比率为3,category_count为1,因此前三行的总和(总数)为240,因为240> 180> 108;对于后两行,比率为5,category_count为2,因此240 + 240 = 480,我想要720 = 240(来自前三行)+ 480(来自后两行)。

我尝试了以下查询

SELECT SUM(total)
FROM (SELECT MAX(total)
      FROM tableName
      GROUP BY rate, category_count) as newTable


它给了我240+ 240,所以我怎么能用category_count(column)给出限制?

最佳答案

您可以通过动态分配变量在MySql中模拟ROW_NUMBER和PARTITION BY

SELECT rate, SUM(total)
FROM
(
  SELECT total, rate, category_count,
         -- Row num counter, resets at each different rate
         @row_num := IF(@prev_rate=rate, @row_num+1, 1) AS RowNum,
         @prev_rate := rate -- Track to see if we are in the same rate
  FROM tableName,
     (SELECT @row_num := 1) x, -- Set initial value for @row_num and @prev_rate
     (SELECT @prev_rate := '') y
  -- Important, must keep rates together, then order by your requirement
  ORDER BY rate ASC, total DESC
) ranked
WHERE ranked.RowNum <= ranked.category_count -- Your requirement of TOP N
GROUP BY rate;


以上收益

Rate Total
3    240
5    480


如果您执行SELECT SUM(total)然后放下GROUP BY,则可以得到720。

SqlFiddle here

编辑

似乎速率定义为(我假设您的样本数据为整数)

`rate` varchar(X) COLLATE utf8_unicode_ci


更改一行:

 (SELECT @prev_rate := '' COLLATE utf8_unicode_ci) y


它将临时跟踪变量设置为与您的列相同的类型

关于mysql - MySQL查询:获取N行的总和,其中N在另一个列值中定义,我们在Stack Overflow上找到一个类似的问题:https://stackoverflow.com/questions/27665175/

10-10 12:27