All, I didn't finish writing this before I posted the previous post. Sigh. Using Rails 1.1.6 AR against MS SQL Server (ODBC mode). I need to do some fancy bulk inserting of data, which I need to ensure is within a transaction. But as I think about it, it isn't clear to me whether what I want to do will work within standard AR. If I do something like this: MyClass.transaction do MyClass.connection.execute("arbitrary SQL") .... MyClass.connection.execute("arbitrary SQL") end will this all use the same connection and actually work within the transaction, even though I'm explicitly doing execute on MyClass.connection _within_ the transaction? I'm concerned that using MyClass.connection implicitly opens a new connection. Can anyone verify whether or not this is the case? If I can't rely on this, I think I should be able to just write up my SQL with explicit SQL-level transaction management within it, and pass that to one call to MyClass.connection.execute. Thanks, Wes
on 2007-06-27 18:55
on 2007-06-27 19:07
On Jun 27, 6:55 pm, Wes Gamble <rails-mailing-l...@andreas-s.net> wrote: > > transaction, even though I'm explicitly doing execute on > Wes What exactly is myclass in this context? Just a AR::Base based class? I execute SQL with ActiveRecord::Base.connection.execute(), and it keeps it within a transaction. If you need to be sure, check out log/ development.log in your project dir, and see what SQL it generates. You should see BEGIN; and COMMIT: around the block of statements.
on 2007-06-27 20:08
MyClass is an AR::Base descendant, correct. My concern is that the execute statements will use a new connection and not be subject to the transaction block in which they are placed. Thanks, Wes