Question

我们可以在SELECT中创建一个返回一行结果的子查询吗？

例如，我必须使用如下所示的查询：

SELECT columnA as aliasA, columnB as aliasB,
(SELECT columnC FROM table2
WHERE clause2) as aliasC,
(SELECT  columnD FROM table2
WHERE clause2) as  aliasD
FROM table1
WHERE clause1

我想改进它，因为它需要太多的时间来执行。有没有办法创建如下的查询：

SELECT columnA as aliasA, columnB as aliasB,
(SELECT columnC, columnD FROM table2
WHERE clause2) as (aliasC, aliasD)
FROM table1
WHERE clause1

由于

注意：更具体地说，我举个例子：我需要在第一季度的每个月获得学生的最佳和最差结果。如果我加入表格，我会为同一个学生提供多行（每个月一个）但我需要：

SELECT firstname, lastname, 
(SELECT MAX(value) FROM results WHERE results.student_id = student.id AND results.date
BETWEEN 'JAN-01' AND 'JAN-31') AS best_result_of_jan,
(SELECT MIN(value) FROM results WHERE results.student_id = student.id AND results.date
BETWEEN 'JAN-01' AND 'JAN-31') AS worst_result_of_jan,
 ..., 
(SELECT MIN(value) FROM results WHERE results.student_id = student.id 
AND results.date BETWEEN 'MAR-01' AND 'MAR-31') AS worst_result_of_mar 
FROM student 
WHERE student.has_maths = 1

Answer 1

修改

SELECT s.firstname, s.lastname, t.best_result_of_jan, t.worst_result_of_mar FROM student s LEFT JOIN ( (SELECT student_id, MAX(CASE WHEN results.date BETWEEN 'JAN-01' AND 'JAN-31' THEN value ELSE NULL END) best_result_of_jan, MIN(CASE WHEN results.date BETWEEN 'MAR-01' AND 'MAR-31' THEN value ELSE NULL END) worst_result_of_mar FROM results GROUP BY student_id ) t ON t.student_id = s.id WHERE s.has_maths = 1

子查询将数据转换为每个学生所需的列。外部查询然后将这些列添加到student表列。

首先，如果子查询返回标量结果（即单行或无结果），则原始查询将起作用。这样可以安全地转换为LEFT JOIN表单，这也可以导致单行（没有笛卡尔积）或没有行（fizzed join）。

SELECT a.columnA as aliasA, a.columnB as aliasB, b.columnC as aliasC, b.columnD as aliasD FROM table1 a LEFT JOIN table2 b ON clause2 WHERE clause1;

确保为第1和第2条中的列添加别名。

注意：此类查询的计划将导致table2处理一次并且columnC / D已解析。您的原始查询可能导致逐行执行子查询，并导致两次启动，因此缺乏性能。

Answer 2

您想要加入，例如：

SELECT table1.columnA as aliasA, table1.columnB as aliasB, 
       table2.columnC as aliasC,
       table2.columnD as aliasD
FROM table1 left outer join
     table2
     on clause2
WHERE clause1

Answer 3

在查询中，如果数据范围参数相同，那么您可以编写一个包含最小和最大字段值的select语句。在目前的情况下，除了其他解决方案，您可以尝试UNION ALL语句并组合两组查询。这可能比子查询方法更快。例如：

Select A.FirstName,
   A.LastName,
   Sum(A.Best_JAN) as BEST_JAN,
   sum(A.Worst_MAR) as WORST_MAR
From (

    Select student.FirstName,
            student.LastName,
            MAX(result.value) as Best_JAN,
            0 as Worst_MAR
      From Student, Results
     Where Student.Studentid = Results.studentid
       and results.date BETWEEN 'JAN-01' AND 'JAN-31'
     Group By student.FirstName, student.LastName

    Union ALL

    Select student.FirstName,
           student.LastName,
           0 as Best_JAN,
           MIN(result.value) as Worst_MAR
      From Student, Results
     Where Student.Studentid = Results.studentid
       And results.date BETWEEN 'MAR-01' AND 'MAR-31'
     Group By student.FirstName, student.LastName

  ) as A

  Group By A.FirstName, A.LastName

在SELECT中创建一个子查询，返回一行结果

3 个答案: