On Apr 27, 10:41 am, Michael Bayer wrote:
>
> then yes, for your case this is exactly the pysqlite bug Daniel mentions:
> http://code.google.com/p/pysqlite/issues/detail?id=21. Pysqlite doesn't open
> the transaction until DML is encountered specifically to reduce file locks.
> This shoul
On Apr 27, 9:37 am, Daniel Holth wrote:
> Is this pysqlite issue about SELECT not starting a transaction
> related?http://code.google.com/p/pysqlite/issues/detail?id=21
Hrmmm... well... that's interesting... it might be related, but maybe
not? I'm not setting the isolation level when I create
On Apr 27, 9:19 am, Michael Bayer wrote:
> SQLite doesn't have support for SELECT..FOR UPDATE, and with_lockmode()
> ultimately has no impact when using SQLite as nothing is rendered.
IDK why sqlite doesn't support a way to elevate the lock on a select
in the middle of a deferred transaction
Yesterday I was working with some code that needed a "select ... for
update" concept to avoid a race condition.
Adding .with_lockmode('update') works a treat on InnoDB and Postgres,
but for sqlite I end up having to sneak in a "if session.bind.name ==
'sqlite'; session.execute('begin immediate tran