Modified the do_begin handling in mssql to use the Cursor not the Connection.

This corrects a problem where we were trying to call execute on the Connection
object instead of against the cursor. This is supported on pyodbc but not in
the DBAPI. Overrode the behavior in pymssql to not do special do_begin
processing on that dialect.

1 files changed, 7 insertions, 2 deletions

diff --git a/lib/sqlalchemy/databases/mssql.py b/lib/sqlalchemy/databases/mssql.py
index 242fc5efb..a3d80c678 100644
--- a/lib/sqlalchemy/databases/mssql.py
+++ b/lib/sqlalchemy/databases/mssql.py
@@ -1056,8 +1056,9 @@ class MSSQLDialect(default.DefaultDialect):
         return newobj
 
     def do_begin(self, connection):
-        connection.execute("SET IMPLICIT_TRANSACTIONS OFF")
-        connection.execute("BEGIN TRANSACTION")
+        cursor = connection.cursor()
+        cursor.execute("SET IMPLICIT_TRANSACTIONS OFF")
+        cursor.execute("BEGIN TRANSACTION")
 
     @base.connection_memoize(('dialect', 'default_schema_name'))
     def get_default_schema_name(self, connection):
@@ -1304,6 +1305,9 @@ class MSSQLDialect_pymssql(MSSQLDialect):
     def is_disconnect(self, e):
         return isinstance(e, self.dbapi.DatabaseError) and "Error 10054" in str(e)
 
+    def do_begin(self, connection):
+        pass
+
 
 class MSSQLDialect_pyodbc(MSSQLDialect):
     supports_sane_rowcount = False
@@ -1430,6 +1434,7 @@ class MSSQLDialect_adodbapi(MSSQLDialect):
     def is_disconnect(self, e):
         return isinstance(e, self.dbapi.adodbapi.DatabaseError) and "'connection failure'" in str(e)
 
+
 dialect_mapping = {
     'pymssql':  MSSQLDialect_pymssql,
     'pyodbc':   MSSQLDialect_pyodbc,