Hello, On Tue, May 20, 2008 at 10:26 PM, D. Richard Hipp <[EMAIL PROTECTED]> wrote: > So it is not at all clear to me whether this behavior is a bug or a > feature. SQLite is doing what the documentation says it ought to do. > The question is, should the specification of what SQLite ought to do > change in order to be less surprising to programmers? And if so, what > would that new specification be?
I wouldn't mind having to call another function, or call it at a different time/location, but if you do an insert, there needs to be *some* way to get the id it just used. :-) Considering triggers: if triggers would have had the side-effect of changing the insert-id, you could still get the "original" insert-id inside the trigger (which would be hard to externalize, but still, you could get your hands on it. If the insert is done in a place you have no control over, you're completely out of luck. As a workaround for now, how is the insert ID generally created (considering a single-threaded app)? select max(id) + 1 from table? - bram _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users