SQLAlchemy 使用 Output 执行存储过程

Question

我正在尝试使用 python 的 sqlalchemy 模块执行存储过程并访问其 output 。

我找到了这篇文章，但无法让它在没有错误的情况下运行。

SQL 存储过程

USE [TestDatabase]
GO

SET ANSI_NULLS ON
GO

SET QUOTED_IDENTIFIER ON
GO

ALTER PROCEDURE [dbo].[testPython]
    @numOne int,
    @numTwo int,
    @numSum int OUTPUT

AS
BEGIN
    LINENO 0
    SET NOCOUNT ON;

    SET @numSum = @numOne + @numTwo
END
GO

Python

import sqlalchemy as db
engine = db.create_engine(f"mssql+pyodbc://{username}:{password}@{dnsname}/{dbname}?driver={driver}")

with engine.connect() as conn:
    outParam = db.sql.outparam("ret_%d" % 0, type_=int)
    result = conn.execute('testPython ?, ?, ? OUTPUT', [1, 2, outParam])
    print(result)

它返回的错误

TypeError：“BindParameter”类型的 object 没有 len()

上述异常是以下异常的直接原因：

SystemError: <class 'pyodbc.Error'> 返回带有错误集的结果

编辑：感谢mkleehammer 的 github ，我已经设法让它与 pyodbc 一起工作。 如果有人知道，我仍然想知道如何使用 sqlalchemy 来做到这一点。

pyodbc 代码

def testsp():
    query = """
    DECLARE @out int;
    EXEC [dbo].[testPython] @numOne = ?, @numTwo = ?, @numSum = @out OUTPUT;
    SELECT @out AS the_output;
    """
    params = (1,2,)

    connection = pyodbc.connect(f'DRIVER={DRIVER_SQL};SERVER={DNSNAME_SQL};DATABASE={DBNAME_SQL};UID={USERNAME_SQL};PWD={PASSWORD_SQL}')
    cursor = connection.cursor()
    cursor.execute(query, params)
    rows = cursor.fetchall()
    while rows:
        print(rows)
        if cursor.nextset():
            rows = cursor.fetchall()
        else:
            rows = None
    cursor.close()
    connection.close()

Answer 1

只需使用简单的 pyodbc 解决方案中相同的匿名代码块的稍微修改版本：

import sqlalchemy as db

engine = db.create_engine("mssql+pyodbc://scott:tiger^5HHH@mssql_199")

query = """\
SET NOCOUNT ON;
DECLARE @out int;
EXEC [dbo].[testPython] @numOne = :p1, @numTwo = :p2, @numSum = @out OUTPUT;
SELECT @out AS the_output;
"""
params = dict(p1=1, p2=2)

with engine.begin() as conn:
    result = conn.execute(db.text(query), params).scalar()
    print(result)  # 3