Question

我有两张桌子：

表t1：

id | date_click
 1 | 2016-02-31 17:17:23
 2 | 2016-03-31 12:11:21
 3 | 2016-03-31 13:13:23

因此，从这张表中我想得到每天的计数字段Id。

为此，我使用下一个查询：

SELECT date_format(date_click, '%Y-%m-%d') as date_click_event
     , COUNT(id) as count_click 
  FROM t1 
 GROUP 
    BY date_click_event 
 ORDER 
    BY date_click_event DESC;

这很好用。

所以下一个表是t2。

id | count | date_sent
 1 |    33 | 2016-02-31 11:12:23
 2 |    22 | 2016-03-31 14:11:22
 3 |    11 | 2016-03-31 13:12:13

要从此表中按日期选择数据，我使用下一个查询：

SELECT date_format(date_sent, '%Y-%m-%d') as date_sent_push
     , SUM(count) as count_sent 
  FROM t2 
 GROUP 
    BY date_sent_push 
 ORDER 
    BY date_sent_push DESC 
 LIMIT 100;

它也很好用。所以我的目的是将这两个查询合并为一个SELECT，然后我可以在php中写一个表Id的表，按表t1的日期计算，并且计数为count表格t2中的字段按日期显示。

当我尝试下一个查询时：

SELECT date_format(t2.date_sent, '%Y-%m-%d') AS date_sent_push
     , SUM(t2.count) as count_sent
     , COUNT(t1.id) as count_click 
  FROM t2
     , t1 
 WHERE date_format(t2.date_sent, '%Y-%m-%d') = date_format(t1.date_click, '%Y-%m-%d') 
 GROUP 
    BY date_sent_push 
 ORDER 
    BY date_sent_push 
  DESC 
 LIMIT 100;

这不行。我做错了什么？

Answer 1

首先，您应该UNION这些结果，然后按天分组并选择聚合字段。您也可以加入这些查询，但如果有两天在两个表之一中错过，则可能会出现问题：

SELECT date_sent_push, 
       MAX(count_click) as count_click,
       MAX(count_sent) as count_sent 
FROM
(SELECT date_format(date_click, '%Y-%m-%d') as date_sent_push
     , COUNT(id) as count_click
     , NULL as count_sent
  FROM t1 
  GROUP BY date_sent_push 
  UNION ALL
  SELECT date_format(date_sent, '%Y-%m-%d') as date_sent_push
     , NULL as count_click
     , SUM(count) as count_sent 
  FROM t2 
  GROUP 
    BY date_sent_push 
) as t3
GROUP BY date_sent_push

SQL fiddle demo

Answer 2

use this code

(SELECT date_format(date_sent, '%Y-%m-%d') as date_sent_push, SUM(count) as count_sent FROM t2 GROUP BY date_sent_push ORDER BY date_sent_push DESC LIMIT 100)
UNION
(SELECT date_format(t2.date_sent, '%Y-%m-%d') AS date_sent_push, SUM(t2.count) as count_sent, COUNT(t1.id) as count_click FROM t2, t1 WHERE date_format(t2.date_sent, '%Y-%m-%d')=date_format(t1.date_click, '%Y-%m-%d') GROUP BY date_sent_push ORDER BY date_sent_push DESC LIMIT 100)

从具有分组日期

2 个答案: