On Thu, Sep 26, 2002 at 11:30:28PM +0200, Bart Lateur wrote: > On Thu, 26 Sep 2002 21:46:16 +0100, Tim Bunce wrote: > > >If that may happen then it's safer to try the insert first and if that > >fails due to a duplicate key then do the update. > > At least MS Access has the problem that an autoincrement field (or > whatever it is called in Access) IS incremented even if the insert > fails. That means, with a 32 bit counter, that sooner or later you might > get in trouble. Well, if you try one update per second, you have 68 > (signed, or 136, with unsigned integers) years or so before it will > overflow, so maybe it's not *that* bad. But it rubs me the wrong way.
MS Access has that effect on a lot of people :-/ > It's a time bomb. If you try 50 inserts per second, you only have a few > years before it happens. Thanks for the info. Tim.