On 2 May 2015, at 12:14pm, Jean-Marie CUAZ <jm.cuaz at orens.fr> wrote:
> Yes, a few years ago on a test db, used for developpment and testing purpose, > "PRAGMA quick_check" reported some anomalies I would not continue to use any system which occasionally caused quick_check -- or more importantly integrity-check -- to fail. If this happens you could lose many rows of a table or even multiple tables. Worse still, pointers may be pointing at the wrong rows, meaning that a future UPDATE or DELETE FROM much affect the wrong rows. In any properly working system even integrity-check (more thorough than quick-check) never fails, and neither should be necessary in anything except a rarely-used maintenance procedure. They are both diagnostic tools for faults, not part of a smoothly running system. If your quick-check reports a fault, you need to replace your hardware or investigate the problem until you've found a cause. If it doesn't, then there's no point in running it and you can save your users from having to wast their time. Simon.