On Sat, Sep 6, 2008 at 7:44 AM, D. Richard Hipp <[EMAIL PROTECTED]> wrote: > On Sep 6, 2008, at 2:50 AM, Scott Hess wrote: >> As a bonus, if the implementation of SQLite changes, SQL itself won't, >> so id will _still_ be a valid integer primary key, even if the >> implementation detail of rowid changes. > > I promise that INTEGER PRIMARY KEY will always be an alias for the > rowid in SQLite. This will not change.
:-). I'm just trying to discourage relying on implementation details when you can easily enough be explicit (in this case, there's no penalty for doing so). -scott _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users