PyODBC return error, but why?

Question

I try execute T-SQL query in two cases (i use Win7, python 3.2, MS SQL Server express 2008,pyodbc for python 3.2):

-case 1:

using MS SQL Server Management Studio i try to execute query:

USE master;CREATE DATABASE Sales ON (NAME = Sales_dat,FILENAME = 'C:\saledat.mdf',        SIZE = 10,    MAXSIZE = 50,    FILEGROWTH = 5 ) LOG ON ( NAME = Sales_log,    FILENAME = 'C:\salelog.ldf',    SIZE = 5MB,    MAXSIZE = 25MB, FILEGROWTH = 5MB );

it return successful result

-case 2:

import pyodbc
cxnn=pyodbc.connect('DSN=SERVER;UID=sa;PWD=password')
cur=cxnn.cursor()
cur.execute("USE master;CREATE DATABASE Sales ON (NAME = Sales_dat,FILENAME = 'C:\saledat.mdf',    SIZE = 10,    MAXSIZE = 50,    FILEGROWTH = 5 ) LOG ON ( NAME = Sales_log,    FILENAME = 'C:\salelog.ldf',    SIZE = 5MB,    MAXSIZE = 25MB, FILEGROWTH = 5MB );")
cxnn.commit()

but after running code i receive error

 pyodbc.Error: ('HY000', 'The driver did not supply an error!')

If i run the code without

cxnn.commit()

i have receive no error. But why?

Answer 1

For reasons I don't quite understand, setting autocommit to be true seems to fix the issue. Please note that escaping the backslashes is still required.

Autocommit can be set in two ways:

cxnn=pyodbc.connect('DSN=SERVER;UID=sa;PWD=password', autocommit=True)

Or:

cxnn=pyodbc.connect('DSN=SERVER;UID=sa;PWD=password')
cxnn.autocommit = True

Once that's done, creating the cursor and executing the query should then run as expected:

cur=cxnn.cursor()
cur.execute("USE master;CREATE DATABASE Sales ON (NAME=Sales_dat, FILENAME='C:\\saledat.mdf', SIZE=10, MAXSIZE=50, FILEGROWTH=5) LOG ON (NAME=Sales_log, FILENAME='C:\\salelog.ldf', SIZE=5MB, MAXSIZE=25MB, FILEGROWTH=5MB );")

Answer 2

In my case I simply needed to upgrade pyodbc for it to work.

pip install --upgrade pyodbc