Question

我有两个表（rescues和counts），在MySQL 5.1数据库中有一对多的关系。 rescues中可能有多个行具有相同的日期时间。在counts中，至少有一行（更多）与rescues相关联（rescues.id = counts.id_rescue）。

我想得到：

“救援”栏目的总和
“计数”中某些列的总和
“计数”中列的总和
按日期对结果进行分组

这就是我正在做的事：http://sqlfiddle.com/#!8/9a290/2/0

我收到了这些结果（错误的sum_volounteers，但更正了sum_alive和sum_dead）：

date        sum_volounteers   sum_alive   sum_dead
2014-02-02  5                 138         18
2014-02-04  4                 53          15

我期待的是：

date        sum_volounteers   sum_alive   sum_dead
2014-02-02  11                138         18
2014-02-04  5                 53          15

Answer 1

我相信你只需SUM()志愿者进行外部查询：

SELECT
r.date,
sum(r.sum_volounteers) AS sum_Volunteers,
sum(c.males + c.females + c.uncertains + c.couples*2) AS sum_alive,
sum(c.roadkills) AS sum_dead
FROM counts AS c
JOIN(
  SELECT
  r2.id AS selected_id,
  sum(r2.volounteers) AS sum_volounteers,
  CAST(r2.start AS DATE) AS date
  FROM rescues AS r2
  GROUP BY selected_id
) AS r ON c.id_rescue = r.selected_id
WHERE r.date >= '2014-02-02' AND r.date <= '2014-02-04'
GROUP BY r.date

演示：SQL Fiddle

由于您只对该日期进行分组，因此MySQL只选择了一个sum_volunteers值从您的子查询返回，而该日期有多个。

Answer 2

编辑：我发现使用此查询时出现问题，如果rescues中有两行以上的行rescues.volounteers中的行号相同，则结果错误。请参阅我的其他答案以获得正确的解决方案。

<击> 感谢其他评论，我发现外部查询需要SUM(DISTINCT r.volounteers)：http://sqlfiddle.com/#!8/9a290/69/0

不需要在子查询中使用SUM()。

SELECT r.date, sum(DISTINCT r.volounteers) AS sum_volounteers, sum(c.males + c.females + c.uncertains + c.couples*2) AS sum_alive, sum(c.roadkills) AS sum_dead FROM counts AS c JOIN( SELECT r2.id AS selected_id, r2.volounteers AS volounteers, CAST(r2.start AS DATE) AS date FROM rescues AS r2 GROUP BY selected_id ) AS r ON c.id_rescue = r.selected_id WHERE r.date >= '2014-02-02' AND r.date <= '2014-02-04' GROUP BY r.date

现在我得到了正确的结果：

date sum_volounteers sum_alive sum_dead 2014-02-02 11 138 18 2014-02-04 5 53 15

<击>

Answer 3

解决方案是在外部查询中选择SUM(r.volounteers) AS sum_volounteers，然后在子查询中获取sum_alive，sum_dead和date_of_sums（使用JOIN子查询）和GROUP BY date_of_sums然后JOIN c.date_of_sums = CAST(r.start AS DATE)

我在rescues中添加的记录在rescues.volounteers列中的编号相同时发布的其他查询失败，这是正确的：

SELECT
   r.id,
   CAST(r.start AS DATE) AS date,
   SUM(r.volounteers) AS sum_volounteers,
   sum_alive,
   sum_dead
FROM rescues AS r
JOIN(
   SELECT
   CAST(r2.start AS DATE) AS date_of_sums,
   sum(c2.males+c2.females+c2.uncertains+c2.couples*2) AS sum_alive,
   sum(roadkills) AS sum_dead
   FROM counts AS c2
   JOIN rescues as r2
   ON r2.id = c2.id_rescue
   GROUP BY date_of_sums
  ) AS c
ON c.date_of_sums = CAST(r.start AS DATE)
WHERE
   CAST(r.start AS DATE) >= '2014-02-02'
   AND CAST(r.start AS DATE) <= '2014-02-04'
GROUP BY CAST(r.start AS DATE)

这是一个演示：

正确查询：http://sqlfiddle.com/#!8/cc1eb/77
~~错误（旧）查询：http://sqlfiddle.com/#!8/efd26/1/0~~

此查询提供正确的结果：

date        sum_volounteers   sum_alive   sum_dead
2014-02-02  11                138         18
2014-02-04  9                 73          23

SQL SELECT SUM来自两个表和GROUP BY日期

3 个答案: