存储过程未使用 sqlalchemy/pyodbc 提交

Question

I am using sqlalchemy/pyodbc to connect to a MS SQL 2012 server. I chose sqlalchemy because of the direct integration with pandas dataframes using .read_sql and .to_sql .

At a high level, my code is:

df = dataframe.read_sql("EXEC sp_getsomedata")
<do some stuff here>
finaldf.to_sql("loader_table", engine,...)

This part works great, very easy to read, etc. The problem is that I have to run a final stored proc to insert the data from the loader table into the live table. Normally, sqlalchemy knows to commit after INSERT/UPDATE/DELETE, but doesn't want to do the commit for me when I run this final stored proc.

After having tried multiple approaches, I see the transaction in the db sitting uncommitted. I know sqlalchemy is very flexible and I am using about 3% of its functionality, what is the simplest way to get this working? I think I need to be using sqlalchemy core and not ORM. I saw examples using sessionmaker, but I think that monopolizes the engine object and doesn't let pandas access it.

connection = engine.connect()
transaction = connection.begin()
connection.execute("EXEC sp_doLoaderStuff")
transaction.commit()
connection.close()

I have tried calling .execute from the connection level, from a cursor level, and even using the .raw_connection() method without success.

connection = engine.raw_connection()
connection.autocommit = True
cursor = connection.cursor()
cursor.execute("EXEC sp_doLoaderStuff")
connection.commit()
connection.close()

Any ideas what I am missing here?

Answer 1

Completely self-inflicted. The correct working code using the raw_connection() method that is working fine is:

connection = engine.raw_connection()
cursor = connection.cursor()
cursor.execute("EXEC sp_doLoaderStuff")
connection.commit()
connection.close()