Question

我正在尝试在我的数据库上执行此查询，该数据库根据第三个表的结果查询2个表。

SELECT *
FROM   ads_user AS u
       INNER JOIN ads_medium AS m
               ON u.id = m.owner_id
WHERE m.id IN (SELECT medium_id,
                                  Count(*) AS count
                           FROM   ads_click
                           WHERE  time > '2017-01-01'
                           GROUP  BY medium_id
                           ORDER  BY count DESC
                           LIMIT  100);

正如您所看到的，我在子查询中使用了两列，count列是整个查询工作所必需的。不幸的是PostgreSql抛出subquery has too many columns错误。

有没有解决方法呢？

Answer 1

只需将计数作为选定列删除，然后将（*）添加到订单中。因此：

SELECT *
FROM   ads_user AS u
       INNER JOIN ads_medium AS m
               ON u.id = m.owner_id
WHERE m.id IN (SELECT medium_id
                           FROM   ads_click
                           WHERE  time > '2017-01-01'
                           GROUP  BY medium_id
                           ORDER  BY count(*) DESC
                           LIMIT  100);

Postgres抱怨是因为你的IN引用了两列，而它应该只有一列。将Count（*）按顺序排列是完全可以的。

如何在子查询中有2列而忽略其中一列？

1 个答案: