如何在 MySQL 的子查询中指定父查询字段？

Question

如何在 MySQL 的子查询中指定父查询字段？

例如：
我用 PHP 编写了一个基本的公告板类型程序。

在数据库中，每个帖子包含：id（PK）和parent_id（父帖子的id）。如果帖子本身是父帖子，则其parent_id 设置为 0。

我正在尝试编写一个 mySQL 查询，该查询将找到每个父帖子以及父帖子所拥有的子帖子数量。

$query = "SELECT id, (
      SELECT COUNT(1) 
      FROM post_table 
      WHERE parent_id = id
) as num_children
FROM post_table
WHERE parent_id = 0";

棘手的部分是第一个 id 不知道它应该引用子查询之外的第二个 id。我知道我可以执行 SELECT id AS id_tmp ，然后在子查询中引用它，但是如果我还想返回 id 并保留“id”作为列名，那么我必须执行一个返回的查询我有两列具有相同的数据（这对我来说似乎很混乱）

$query = "SELECT id, id AS id_tmp, 
            (SELECT COUNT(1)
            FROM post_table
            WHERE parent_id = id_tmp) as num_children
         FROM post_table
         WHERE parent_id = 0";

这种混乱的方式工作得很好，但我觉得有机会在这里学习一些东西，所以我想我应该发布这个问题。

Answer 1

怎么样：

$query = "SELECT p1.id, 
                 (SELECT COUNT(1) 
                    FROM post_table p2 
                   WHERE p2.parent_id = p1.id) as num_children
            FROM post_table p1
           WHERE p1.parent_id = 0";

或者如果你在 p1.id 上添加一个别名，你可能会说：

$query = "SELECT p1.id as p1_id, 
                 (SELECT COUNT(1) 
                    FROM post_table p2 
                   WHERE p2.parent_id = p1.id) as num_children
            FROM post_table p1
           WHERE p1.parent_id = 0";

Answer 2

您可以尝试类似的操作

SELECT  pt.id,
        CountTable.Cnt
FROM    post_table pt LEFT JOIN
        (
            SELECT  parent_id,
                    COUNT(1) Cnt
            FROM    post_table
            WHERE   parent_id <> 0
            GROUP BY parent_id
        ) CountTable ON pt.id = CountTable.parent_id
WHERE   pt.parent_id = 0

要回到您的示例，请在子选择中使用主表的别名

SELECT  pt.id,
        (SELECT COUNT(1) FROM post_table WHERE parent_id = pt.id) 
FROM    post_table pt
WHERE   pt.parent_id = 0

Answer 3

为表指定唯一的名称：

$query = "SELECT a.id, (SELECT COUNT(1) FROM post_table b WHERE parent_id = a.id) as num_children FROM post_table a WHERE a.parent_id = 0";

Answer 4

以下语法适用于 Oracle。你能测试一下在 MYSQL 中是否也同样有效吗？
在 Oracle 中称为标量子查询。

如果您两次使用同一个表，您只需要为两个表添加不同的别名即可区分它们。

sql> select empno,
  2         (select dname from dept where deptno = emp.deptno) dname
  3    from emp 
  4    where empno = 7369;

     EMPNO DNAME
---------- --------------
      7369 RESEARCH

sql> select parent.empno,
  2         (select mgr from emp where empno = parent.empno) mgr
  3    from emp parent
  4    where empno = 7876;

     EMPNO        MGR
---------- ----------
      7876       7788

Answer 5

谢谢唐。我有一个如下所示的嵌套查询，其中的 WHERE 子句无法确定别名 v1。这是不起作用的代码：

Select 
    teamid,
    teamname
FROM
    team as t1
INNER JOIN (
    SELECT 
        venue_id, 
        venue_scores, 
        venue_name 
    FROM venue 
    WHERE venue_scores = (
        SELECT 
            MAX(venue_scores) 
        FROM venue as v2 
        WHERE v2.venue_id = v1.venue_id      /* this where clause wasn't working */
    ) as v1    /* v1 alias already present here */
);

因此，我只是在 JOIN 中再次添加了别名 v1 。这使它发挥作用。

Select 
    teamid,
    teamname
FROM
    team as t1
INNER JOIN (
    SELECT 
        venue_id, 
        venue_scores, 
        venue_name 
    FROM venue as v1              /* added alias v1 here again */
    WHERE venue_scores = (
        SELECT 
            MAX(venue_scores) 
        FROM venue as v2 
        WHERE v2.venue_id = v1.venue_id   /* Now this works!! */
    ) as v1     /* v1 alias already present here */
);

希望这对某人有帮助。

Answer 6

MySQL 8 中子查询中的父查询字段。

我使用嵌套查询根据 tblgamescores 中的用户名选择游戏分数。

SELECT 
    GameScoresID, 
    (SELECT Username FROM tblaccounts WHERE AccountID = FromAccountID) AS FromUsername, 
    (SELECT Username FROM tblaccounts WHERE AccountID = ToAccountID) AS ToUsername,
    (SELECT Username FROM tblaccounts WHERE AccountID = WinAccountID) AS WinUsername,
    (SELECT Username FROM tblaccounts WHERE AccountID = LossAccountID) AS LossUsername,
    FromUserScore,
    ToUserScore 
FROM tblgamescores a 
WHERE FromAccountID = (SELECT AccountID FROM tblaccounts WHERE Username = "MHamzaRajput");