Question

SELECT events.id, events.name, events.code, event_statuses.name as event_status_name
 , regions.name as event_region_name, event_locations.name as event_location_name
 , events.date_1, events.date_2
 , count(r.id) as regs 
FROM events 
 INNER JOIN event_locations ON event_locations.id = events.event_location_id 
 INNER JOIN event_statuses ON event_statuses.id = events.event_status_id
 LEFT OUTER JOIN regions ON regions.id = events.region_id 
 LEFT OUTER JOIN registrations as r ON events.id = r.event_id and r.registration_status_id = 4 
WHERE 1 GROUP BY events.id ORDER BY date_1 DESC LIMIT 0, 50

如您所见，我的查询具有聚合函数，它在按事件分组的注册上执行。

但是，我希望在此查询中有另一个GROUP BY，这个没有任何其他聚合函数。这里的要点是我希望按区域对事件列表进行分组 - 现在对它们排序的内容很重要。如果它们按regs排序 - 我会将来自波士顿的所有事件聚集在一起，然后按照regs进行排序。因此需要另一个Group By。这就是我提出的，但这也不起作用：

SELECT events.id, events.name, events.code, event_statuses.name as event_status_name
 ,regions.name as event_region_name, event_locations.name as event_location_name
 , events.date_1, events.date_2, r.regs
 FROM events 
 INNER JOIN event_locations ON event_locations.id = events.event_location_id 
 INNER JOIN event_statuses ON event_statuses.id = events.event_status_id
 LEFT OUTER JOIN regions ON regions.id = events.region_id 
 LEFT OUTER JOIN 
        (Select event_id, count(id) as regs 
        from registrations where registration_status_id = 4 
        group by event_id) r on events.id = r.event_id
 WHERE 1 GROUP BY events.region_id, events.id ORDER BY date_1 DESC LIMIT 0, 50

我仍然不能按地区将它们分组。

Answer 1

我希望按区域对事件列表进行分组 - 现在对它们排序的内容很重要。如果它们按照regs排序 - 我会将来自波士顿的所有事件聚集在一起，然后按照regs进行排序。

你想要ORDER BY，而不是GROUP BY。您可以在ORDER BY子句中的所有其他字段之前添加要“聚集在一起”的字段。

ORDER BY region_id, regs

Answer 2

这是修改后的版本

SELECT
  events.id,
  events.region_id,
  events.name,
  events.code,
  event_statuses.name  as event_status_name,
  regions.name         as event_region_name,
  event_locations.name as event_location_name,
  events.date_1,
  events.date_2,
  r.regs
FROM events
  INNER JOIN event_locations
    ON event_locations.id = events.event_location_id
  INNER JOIN event_statuses
    ON event_statuses.id = events.event_status_id
  LEFT OUTER JOIN regions
    ON regions.id = events.region_id
  LEFT OUTER JOIN (Select
                 event_id,
                 count(id)             as regs
               from registrations
               where registration_status_id = 4
               group by event_id) r
    on events.id = r.event_id
WHERE 1
GROUP BY events.region_id, events.id
ORDER BY date_1 DESC
LIMIT 0, 50

试试这个我只选择了events.region_id列

Answer 3

如果你想“聚集在一起”，那么听起来就像你想要使用并汇总为regs。

SELECT 
  events.id, 
  events.name, 
  events.code, 
  event_statuses.name as event_status_name,
  regions.name as event_region_name, 
  event_locations.name as event_location_name, 
  events.date_1, 
  events.date_2, 
  SUM(r.regs) AS regs
....
GROUP BY events.region_id, events.id ORDER BY date_1 DESC LIMIT 0, 50

查询中有两个分组，一个没有聚合函数。

3 个答案: