Обсуждение: Checking of constraints via subqueries?

Hallo all,

in part VI (the reference), the online manual for 8.0.1 states for command
CREATE DOMAIN for parameter
    CHECK (expression)
that currently
     "...CHECK expressions cannot contain subqueries nor refer to
     variables other than VALUE."

So, is this an item for the TODO list?  I can't find it there.  Or do I
overlook something?
This is also stated in the page about CREATE TABLE.

Best regards,
    Wolfgang

On Thu, Feb 17, 2005 at 04:37:43PM +0100, Wolfgang Drotschmann wrote:

> "...CHECK expressions cannot contain subqueries nor refer to
>  variables other than VALUE."

In experiments I've defined a domain's CHECK expression to call a
function that makes queries.  It worked in simple tests; I don't
know if it'll work in the general case.

--
Michael Fuhr
http://www.fuhr.org/~mfuhr/

On Thu, Feb 17, 2005 at 11:03:58AM -0700, Michael Fuhr wrote:
> On Thu, Feb 17, 2005 at 04:37:43PM +0100, Wolfgang Drotschmann wrote:
>
> > "...CHECK expressions cannot contain subqueries nor refer to
> >  variables other than VALUE."
>
> In experiments I've defined a domain's CHECK expression to call a
> function that makes queries.  It worked in simple tests; I don't
> know if it'll work in the general case.

The problem is basically that modifying a table elsewhere could have
the effect of invalidating the entries in the table being checked. So
while having the CHECK constraint call a function will work, it will
not guarentee that it will be checked whenever the result of the
subquery changes. The simplest example I can think of would be:

CREATE TABLE blah (
  ...
  field INT4 CHECK( field > (SELECT value FROM configtable WHERE key = 'field_limit') ),
  ...
);

If you hide the comparison within a function PostgreSQL will let you,
but it won't guarentee that it will always be true. For example, if you
modify the configtable you could invalidate the constraint for many
rows at once, yet PostgreSQL has no way of knowing that.

This case is simple but as the subqueries become more complex, it gets
harder to see a shortcut. Eventually you'd have to resort to rechecking
every row of table blah every time any query modifies any dependant
table.

Hope this helps,
--
Martijn van Oosterhout   <kleptog@svana.org>   http://svana.org/kleptog/
> Patent. n. Genius is 5% inspiration and 95% perspiration. A patent is a
> tool for doing 5% of the work and then sitting around waiting for someone
> else to do the other 95% so you can sue them.

On Thu, Feb 17, 2005 at 07:31:42PM +0100, Martijn van Oosterhout wrote:
> On Thu, Feb 17, 2005 at 11:03:58AM -0700, Michael Fuhr wrote:
> >
> > In experiments I've defined a domain's CHECK expression to call a
> > function that makes queries.  It worked in simple tests; I don't
> > know if it'll work in the general case.
>
> The problem is basically that modifying a table elsewhere could have
> the effect of invalidating the entries in the table being checked.

Yeah, that's the kind of thing that concerned me.  I meant to ask
the OP the usual question:

What problem are you trying to solve?

--
Michael Fuhr
http://www.fuhr.org/~mfuhr/

Michael Fuhr wrote:
> On Thu, Feb 17, 2005 at 07:31:42PM +0100, Martijn van Oosterhout wrote:
>
>>On Thu, Feb 17, 2005 at 11:03:58AM -0700, Michael Fuhr wrote:
>>
>>>In experiments I've defined a domain's CHECK expression to call a
>>>function that makes queries.  It worked in simple tests; I don't
>>>know if it'll work in the general case.
>>
>>The problem is basically that modifying a table elsewhere could have
>>the effect of invalidating the entries in the table being checked.
>
>
> Yeah, that's the kind of thing that concerned me.  I meant to ask
> the OP the usual question:
>
> What problem are you trying to solve?

No problem actually.

Realising a constraint via a CHECK (<subquery>) is a decision made by the
developer.  So the developer has to make sure that this fits in the whole
picture and works the intended way.
I'm thinking about the feature, and not about whether the use of this feature
is dangerous or not.
My starting point was the thought of using this for checking in CREATE DOMAIN.
I see that the ubiquitous triggers could help (in an indirect way, being
called with every INSERT or UPDATE of a row with that domain, but is this
generic?).  The other (subquery) option looks quite more elegant (and
generic).  And that may be the reason why it is mentioned in the docs: it
seems to be an elegant and natural way the check a domain.

Best regards,
    Wolfgang