使用 ROW_NUMBER() window 函数选择行

Question

我目前正在为工作中的流程编写查询。我试图通过应用 ROW_NUMBER() 函数来获取特定行，然后根据行定位选择行。但是，我得到以下 error:

SQL compilation error: Window function [ROW_NUMBER() OVER (ORDER BY EMPLOYEE.SALARY_GRADE_ID ASC NULLS LAST)] appears outside of SELECT, QUALIFY, and ORDER BY clauses.

查询：

SELECT Employee.Salary_Grade_Id, SUM(Salary_Grades.Grade_Amount) AS total, ROW_NUMBER() OVER(ORDER 
BY Employee.Salary_Grade_Id) AS rowCol FROM Employee, Salary_Grades 
WHERE (Employee.Salary_Grade_Id = Salary_Grades.Grade_Id AND rowCol = 1) GROUP BY 
Employee.Salary_Grade_Id;

没有意义的是，当我删除 AND rowCol = 1 时，查询没有错误。但是有了它，我得到了错误。这是怎么回事？

Answer 1

您的问题是 rowCol 是 window 函数 (ROW_NUMBER()) 的别名，它们不能出现在 WHERE 子句中。您可以改用 QUALIFY 子句：

SELECT Employee.Salary_Grade_Id, 
       SUM(Salary_Grades.Grade_Amount) AS total, 
       ROW_NUMBER() OVER(ORDER BY Employee.Salary_Grade_Id) AS rowCol 
FROM Employee, Salary_Grades 
WHERE (Employee.Salary_Grade_Id = Salary_Grades.Grade_Id)
GROUP BY Employee.Salary_Grade_Id
QUALIFY rowCol = 1

请注意，您应该使用明确的 JOIN 语法并将查询重写为

SELECT Employee.Salary_Grade_Id, 
       SUM(Salary_Grades.Grade_Amount) AS total, 
       ROW_NUMBER() OVER(ORDER BY Employee.Salary_Grade_Id) AS rowCol 
FROM Employee
JOIN Salary_Grades ON Employee.Salary_Grade_Id = Salary_Grades.Grade_Id
GROUP BY Employee.Salary_Grade_Id
QUALIFY rowCol = 1

Answer 2

WHERE 子句在窗口之前应用，因此您不能在 WHERE 子句中引用窗口函数。您需要使用子查询：

SELECT 
    grade_id, 
    sum(grade_amount) as total
FROM (
    SELECT
        Employee.Salary_Grade_Id as grade_id, 
        Salary_Grades.Grade_Amount AS grade_amount, 
        ROW_NUMBER() OVER(ORDER BY Employee.Salary_Grade_Id) AS rowCol 
    FROM Employee JOIN Salary_Grades ON Employee.Salary_Grade_Id = Salary_Grades.Grade_Id) x
WHERE  rowCol = 1
GROUP BY grade_id;

我还将您的连接谓词移到了 FROM 子句中，作为避免在谓词 Employee.Salary_Grade_Id = Salary_Grades.Grade_Id 被错误删除时意外笛卡尔连接的最佳实践。

使用 ROW_NUMBER() window 函数选择行

Selecting row with ROW_NUMBER() window function

sql

snowflake-cloud-data-platform