[pysqlite] How do I use pysqlite in a multi-threading env.?
In my wxPython-app a part of it gathers data, when a button is pressed, and stores it into a db. The GUI part should display the stuff being stored in the db. When both parts work on the same connection, I get "SQL statements in progress errors". Seems ok to me, you can't do that. So, next step: Both parts get a separate connection. Now I get "Database locked" errors. Hmm, yes, it's GUI stuff after all, being all in the same thread. So, next step: I put the data gathering part into a real thread, that's started, wehn the "Scan" button is pressed. This way the "Database locked"- errors should got away. But now I get "SQLite objects created in a thread can only be used in that same thread.The object was created in thread id 3576 and this is thread id 1040". Hmm, true, I want to display db content, that was stored by an other thread. But that's what multi-threading capabilities are for! I must be missing something here. Anyway, I'm a bit lost now, really. So, how are you doing such stuff with pysqlite? Kind regards Franz GEIGER -- http://mail.python.org/mailman/listinfo/python-list
Re: [pysqlite] How do I use pysqlite in a multi-threading env.?
On Fri, May 20, 2005 at 08:42:54AM +0200, F. GEIGER wrote: > In my wxPython-app a part of it gathers data, when a button is pressed, and > stores it into a db. > > The GUI part should display the stuff being stored in the db. > > When both parts work on the same connection, I get "SQL statements in > progress errors". Seems ok to me, you can't do that. You get that error message with which pysqlite call? > So, next step: Both parts get a separate connection. Now I get "Database > locked" errors. Hmm, yes, it's GUI stuff after all, being all in the same > thread. One connection did not commit() or rollback() within the default timeout (5 seconds). SQLite is not a good solution if you need long transactions, because other SQLite connections are blocked, then. > So, next step: I put the data gathering part into a real thread, that's > started, wehn the "Scan" button is pressed. This way the "Database locked"- > errors should got away. But now I get "SQLite objects created in a thread > can only be used in that same thread.The object was created in thread id > 3576 and this is thread id 1040". Hmm, true, I want to display db content, > that was stored by an other thread. But that's what multi-threading > capabilities are for! Yes, but pysqlite's threadsafety level is 1, that means you cannot share a connection object among multiple threads. In order to ensure that and to avoid that you get random and unexplainable crashes, I implemented this warning. > I must be missing something here. Anyway, I'm a bit lost now, really. So, > how are you doing such stuff with pysqlite? Use multiple connections and commit/rollback early. HTH, -- Gerhard -- Gerhard Häring - [EMAIL PROTECTED] - Python, web & database development signature.asc Description: Digital signature -- http://mail.python.org/mailman/listinfo/python-list
Re: [pysqlite] How do I use pysqlite in a multi-threading env.?
"Gerhard Haering" <[EMAIL PROTECTED]> schrieb im Newsbeitrag news:[EMAIL PROTECTED] I've completely rewritten the db handling stuff, which formerly did too much behind the scenes (explicit is better than implicit...). Now everything looks much better. No unexpected errors and - most important - locking works as expected. I have no idea yet, what went wrong with the former (probably more pythonic) solution. It always did it with MySQL as backend. Anyway, I'm quite happy now, that I can work on a db, that is incorporated into the app itself. Thanks for your help, Gerhard! And, sorry for the troubles... Kind regards Franz GEIGER -- http://mail.python.org/mailman/listinfo/python-list
