from contextlib import closing
def init_db():
with closing(connect_db()) as db:
with app.open_resource(\'schema.sql\') as f:
db.cursor().execute
The only thing done by the with
statement is to call __enter__
method before entering its block and __exit__
method before exiting it.
If those methods are not defined the with
statement won't work as you may expect. I don't know what is the return type of connect_db
, but I guess that it could be many different things from different third-party libraries. So, your code without closing
will probably work in many (all?) cases, but you never know what can be returned by connect_db
.
Yes, you should be using context.closing()
; your own version does something different entirely.
The with
statement lets a context manager know when a block of code is entered and exited; on exit the context manager is also given access to the exception, if one occurred. File objects use this to automatically close the file when the block is exited.
The connect_db()
function from the tutorial returns a sqlit3 connection object, which can indeed be used as a context manager. However, the connection.__exit__()
method doesn't close the connection, it commits the transaction on a successful completion, or aborts it when there is an exception.
The contextlib.closing()
context manager on the other hand, calls the connection.close()
method on the connection. This is something entirely different.
So, your second snippet may work, but does something different. The tutorial code closes the connection, your version commits a transaction. You are already calling db.commit()
, so the action is actually redundant provided no exceptions are raised.
You could use the connection as a context manager again to have the automatic transaction handling behaviour:
def init_db():
with closing(connect_db()) as db:
with app.open_resource('schema.sql') as f, db:
db.cursor().executescript(f.read())
Note the , db
on the second with
line, ensuring that the db.__exit__()
method is called when the block exits.