pyodbc

Question

所以我试图用 group by 子句中的参数编写一个 SQL 查询并让 pyodbc 执行它；我有以下架构：

如果 sqlfiddle 失败，

CREATE TABLE Persons (
    Personid int IDENTITY(1,1) PRIMARY KEY,
    Name varchar(255) NOT NULL,
    Birthday datetime
);

INSERT INTO PERSONS (NAME, BIRTHDAY)
VALUES
('a','20220101'),
('b','20220102'),
('c','20220103'),
('d','20220104'),
('e','20220105'),
('f','20220106'),
('g','20220108'),
('h','20220110'),
('i','20220111'),
('j','20220112'),
('k','20220113'),
('l','20220114'),
('m','20220115')

我以下面的查询为例，它是有效的 SQL:

select
  COUNT(*)
  ,dateadd(week, datediff(week,0, birthday), 0)
 from Persons

group by 
dateadd(week, datediff(week,0, birthday), 0)
order by
dateadd(week, datediff(week,0, birthday), 0)

此查询将按周对用户的生日进行分组。这只是一个人为的例子。我有与此类似的真实数据。

我正在尝试编写一个 python 函数，它可以获取数据并按周对它们进行分组。我希望这个函数能够决定一周的哪一天是一周的开始。我有以下功能：

import pyodbc
def TestSQLServerDB2(dayOfWeekStart=0):
    """Tests the query.

    :param: dayOfWeekStart: Int. 0 = Monday, 1 = Tuesday, ... 6 = Saturday
    """

    hostname = 'DESKTOPHOST'
    database_instance = "test"
    db_conn = pyodbc.connect('Trusted_Connection=yes;' + r"DRIVER=" + "{SQL SERVER}" +
                    ";SERVER=" + hostname + ";DATABASE=" +
                    database_instance + ";")


    targetWeekday = dayOfWeekStart % 7

    sql = '''
        select
        COUNT(*)
        ,dateadd(week, datediff(week,0, birthday), ?)
        from Persons

        group by 
            dateadd(week, datediff(week,0, birthday), ?)
        order by
            dateadd(week, datediff(week,0, birthday), ?)
    '''

    params = (targetWeekday ,targetWeekday ,targetWeekday )
    
    cur = db_conn.cursor()
    cur.execute(sql, params)
    
    print(cur.fetchall())

运行这个函数会产生如下错误：

pyodbc.ProgrammingError: ('42000', "[42000] [Microsoft][ODBC SQL Server Driver][SQL Server]Column 'Persons.Birthday' is invalid in the select list because it is not contained in either an aggregate function or the GROUP BY clause. (8120) (SQLExecDirectW); [42000] [Microsoft][ODBC SQL Server Driver][SQL Server]Statement(s) could not be prepared. (8180)")

在执行之前对 SQL 查询进行字符串替换对我来说已经足够简单了，我不会遇到任何问题；我想知道的是 SQL 参数化语句是否基本上仅限于 WHERE 子句中的值比较？这是有意设计的吗？

谢谢

Answer 1

问题是您基本上将参数传递到查询中三次，而 GROUP BY 和 SELECT 使用不同的参数。尽管 you 知道它们相同，但服务器不知道。实际上，您的查询是：

select
  COUNT(*)
  ,dateadd(week, datediff(week,0, birthday), @p0)
 from Persons

group by 
dateadd(week, datediff(week,0, birthday), @p1)
order by
dateadd(week, datediff(week,0, birthday), @p2)

很明显为什么它不起作用。

解决方案是对每个使用相同的值。

不幸的是，你不能在 pyodbc 中使用命名参数，所以那是不可能的。
您可以使用变量，但这会禁用参数嗅探，这可能是不可取的。
您可以将整个内容放在嵌套的 subquery/derived table 中。但是嵌套很乱，over-complicates查询。
相反，将值放在 CROSS APPLY (VALUES 中，这意味着您可以在查询中的任何地方重复使用它

select
  COUNT(*)
  ,v.birthweek
 from Persons
 cross apply (values (
    dateadd(week, datediff(week,0, birthday), ?)
 ) ) v(birthweek)
group by 
  v.birthweek
order by
  v.birthweek;

db<>fiddle

无论如何，这都是避免重复计算的明智解决方案。

pyodbc - group/order 子句中的参数 - 限制？

pyodbc - parameters in group/order clause - Limitation?

python

sql

sql-server