On 25 Jun 2013, at 1:41am, jhnlmn <[email protected]> wrote: > Using trigger during insert slows down insert by about 20%. > This is not much, but unnecessary as long as AUTOINCREMENT > does not reach the maximum value. > As other people mentioned, this is unlikely to happen > in our lifetime. But if it will, then this trigger/second table approach > may be a solution.
It won’t. 2^64 is about 2e19. Let’s suppose you could find a storage system that would run 24/7 and not break down for a long long time. Let’s suppose you can insert a million rows a second. It would take about half a million years of continuous operation to overflow the AUTOINC number. I bet you a slap up dinner at the best hotel in the world if you still care about this issue in even a hundred thousand years time. Simon. _______________________________________________ sqlite-users mailing list [email protected] http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
