Ah!
I didn't know that!
Michael, thank you so much! No, I don't have anything sharing these
sessions, they are rules in the back end so they should (unless someone
tells me differently) AFAIK be in their own threads.
Top man!
Brad
Michael Fuhr wrote:
>On Wed, Sep 22, 2004 at 04:30:06PM +0100, Bradley Kieser wrote:
>
>
>>Michael Fuhr wrote:
>>
>>
>>
>>>CREATE RULE v_ins AS ON INSERT TO v DO INSTEAD (
>>>INSERT INTO a (xxx) VALUES (NEW.xxx);
>>>INSERT INTO b (col1, col2, col3) VALUES (NEW.col1, NEW.col2,
>>>currval('a_id_seq'))
>>>);
>>>
>>>
>>But what if another insert happens in the mean time? Then the sequence
>>would have advanced that the data will be scrambled. |Given that this
>>could be a very active dataset, that situation is almost certain to occur.
>>
>>
>
>currval() returns the last value returned by nextval() in *this session*.
>
>http://www.postgresql.org/docs/7.4/static/functions-sequence.html
>http://www.postgresql.org/docs/faqs/FAQ.html#4.15.3
>
>Do you have multiple processes or threads sharing the same connection?
>Perhaps one of the developers can comment on the possibility of a race
>condition in that case.
>
>
>