Hiroshi Inoue wrote:
> Richard Huxton wrote:
>> Heikki Linnakangas wrote:
>>> The problem is that the new tuple version is checked only against the
>>> condition in the update rule, id=OLD.id, but not the condition in the
>>> original update-claus, dt='a'.
>>>
>>> Yeah, that's confusing :(.
>>
>> Bit more than just normal rule confusion I'd say. Try the following
>> two statements in parallel (assuming you've just run the previous):
>>
>> UPDATE test SET dt='c';
>> UPDATE test SET dt='x' FROM test t2 WHERE test.id=t2.id AND t2.dt='b';
>>
>> This isn't a problem with the view mechanism - it's a problem with
>> re-checking clauses involving subqueries or joins I'd guess.
>
> I don't understand the PostgreSQL specific *FROM* clause correctly.
> Currently the relations in the *FROM* clause seem to be read only
> and UPDATE operations seem to acquire no tuple level lock on them.
Yes, the above query is equivalent to:
UPDATE test SET dt='x' WHERE id IN (SELECT id FROM test WHERE dt='b');
There are some expressions more naturally expressed as a set of where
conditions though, and I think the "FROM" is just to provide a place to
name them.
The FROM form seemed to be the more natural match to the plan your view
was generating - I'm not sure which the plan transformation process
produces.
-- Richard Huxton Archonet Ltd