Doug McNaught wrote:
>Dmitry Tkach <dmitry@openratings.com> writes:
>
>
>
>>select setval('answer_id_seq', id) from answer order by id desc limit 1;
>>
>>Now, for some reason this reports a correct value, but what actually
>>gets set is wrong!
>>
>>
>
>I'm guessing that 'setval' is getting called more than once here.
>Your 'LIMIT 1' controls how many rows are returned to the client, but
>the server is probably generating more rows internally. So this is
>just wrong, and
>
>
You are right ! That's it! It does get called twice.
There is even a nice comment in ExecLimit():
* NOTE: when scanning forwards, we must fetch one tuple beyond the
* COUNT limit before we can return NULL, else the subplan won't
* be properly positioned to start going backwards. Hence test
* here is for position > netlimit not position >= netlimit.
Whatever that means, that's what was causing my problems...
Thanks!
Dima
