Re: BUG #6612: Functions can be called inside CHECK statements
От
Tom Lane
Тема
Re: BUG #6612: Functions can be called inside CHECK statements
Дата
Msg-id
6552.1335368761@sss.pgh.pa.us
Ответ на
Re: BUG #6612: Functions can be called inside CHECK
statements (Kevin Grittner)
Список
Дерево обсуждения
BUG #6612: Functions can be called inside CHECK statements atrigent@ccs.neu.edu
Re: BUG #6612: Functions can be called inside CHECK statements hubert depesz lubaczewski <depesz@depesz.com>
Re: BUG #6612: Functions can be called inside CHECK statements Greg Stark <stark@mit.edu>
Re: BUG #6612: Functions can be called inside CHECK
statements "Kevin Grittner" <Kevin.Grittner@wicourts.gov>
Re: BUG #6612: Functions can be called inside CHECK statements Tom Lane <tgl@sss.pgh.pa.us>
Re: BUG #6612: Functions can be called inside CHECK statements Greg Stark <stark@mit.edu>
Re: BUG #6612: Functions can be called inside CHECK statements Tom Lane <tgl@sss.pgh.pa.us>
"Kevin Grittner" writes: > Greg Stark wrote: >> Only IMMUTABLE functions can be used in CHECK constraints. > It doesn't look like that to me: No, we have never enforced that. IIRC the idea has been discussed, but we thought that adding the restriction would break too many existing applications. > A CHECK constraint using a volatile function is potentially valid > and useful, IMO. Think about a column which is supposed to record > the moment of an event which has occurred. It could make sense to > ensure that the timestamptz value is < now(); On the other hand, an > index entry based on now() is clearly a problem. This example is actually stable not volatile, but if for some reason you wanted to use clock_timestamp() then it would be volatile. Probably a more interesting question is whether it'd ever be sane to use a function with side-effects in a check constraint. I find it hard to visualize a case where it wouldn't be saner to put the actions in a trigger, but that doesn't mean someone else might not wish to do it. In practice, the times when check constraints are checked are predictable enough that you should be able to get away with abusing the system like that, if you wished. regards, tom lane
В списке pgsql-bugs по дате отправления
От: Greg Stark
Дата: