左连接中的子查询

Question

LEFT JOIN团队在teams.id =(SELECT team_id FROM auth_users WHERE id = offers.user_id)

给我团队表的所有列,但将所有列转换为NULL.

LEFT JOIN teams ON teams.id = 1

奇迹般有效

当我做

SELECT (SELECT team_id FROM auth_users WHERE id = offers.user_id) as team_id

team_id将为1.

由于一些奇怪的原因,它在内部不起作用JOIN.

完整查询:

SELECT projects.id, projects.title as title, winner_id, projects.user_id as user_id,   until, pages, types.title as type, types.id as type_id, projects.id as id, offers.price as price, offers.delivery_date as delivery_date, teams.*,
(SELECT COUNT(id) FROM comments WHERE comments.project_id = projects.id AND comments.private = 1) as comments,
(SELECT COUNT(id) FROM uploads WHERE uploads.project_id = projects.id) as files,
(SELECT country FROM auth_users WHERE auth_users.id = offers.user_id) as baser_country,
(SELECT business FROM auth_users WHERE auth_users.id = offers.user_id) as baser_business,
(SELECT CONCAT(firstname, ' ', lastname) FROM auth_users WHERE auth_users.id = offers.user_id) as baser_name,
(SELECT team_id FROM auth_users WHERE id = offers.user_id) as team_id,
(SELECT country FROM auth_users WHERE auth_users.id = projects.user_id) as customer_country,
(SELECT business FROM auth_users WHERE auth_users.id = projects.user_id) as customer_business
FROM projects
JOIN types ON projects.type_id = types.id
LEFT JOIN offers ON projects.id = offers.project_id
LEFT JOIN teams ON teams.id = (SELECT team_id FROM auth_users WHERE id = offers.user_id)
WHERE projects.user_id = 1 AND winner_id != 0 AND uploaded = 1
GROUP BY projects.id
ORDER BY projects.id DESC
LIMIT 3

提前致谢!

Answer 1

首先,你应该摆脱所有这些子查询.通过正确的连接,它们都不是必需的.它们使查询非常拥挤,并且它们对查询性能没有贡献.

我最好的理解你的查询应该是这样的.

SELECT 
  p.id                 AS project_id, 
  MIN(p.title)         AS project_title, 
  MIN(winner_id)       AS winner_id, 
  MIN(p.user_id)       AS project_user_id,
  MIN(until)           AS until, 
  MIN(pages)           AS pages, 
  MIN(t.id)            AS type_id, 
  MIN(t.title)         AS type_title, 
  MIN(o.price)         AS offer_price, 
  MIN(o.delivery_date) AS offer_delivery_date, 
  -- m.*, <-- this should be avoided, especially in a grouped query
  COUNT(c.id)          AS count_comments,
  COUNT(u.id)          AS count_files,
  MIN(ao.country)      AS baser_country,
  MIN(ao.business)     AS baser_business,
  MIN(CONCAT(ao.firstname, ' ', ao.lastname)) AS baser_name,
  MIN(ao.team_id)      AS baser_team_id,
  MIN(ap.country)      AS customer_country,
  MIN(ap.business)     AS customer_business
FROM 
  projects p
  INNER JOIN types       t ON p.type_id    = t.id
  LEFT  JOIN offers      o ON o.project_id = p.id
  LEFT  JOIN comments    c ON c.project_id = p.id AND comments.private = 1
  LEFT  JOIN uploads     u ON u.project_id = p.id
  LEFT  JOIN auth_users ao ON ao.id = o.user_id
  LEFT  JOIN auth_users ap ON ap.id = p.user_id
  LEFT  JOIN teams       m ON m.id  = o.team_id
WHERE 
  p.user_id = 1 
  AND winner_id != 0 
  AND uploaded = 1
GROUP BY 
  p.id
ORDER BY 
  p.id DESC
LIMIT 3

接下来,您永远不应该编写一个分组查询,其中输出中的字段既不是分组也不是聚合.我在MIN()查询的所有流氓未分组字段中使用了聚合函数.

我建议使用短表别名,我建议在任何地方使用表别名:作为第三方读者,当谈到表" winner_id"来自的问题时,我完全迷失了.