Tom Lane wrote:
> Christian Rank <christian.rank@rz.uni-passau.de> writes:
>
>>... according to the docs, the validity of a constraint should be
>>checked after each statement unless this behaviour is altered with a SET
>>CONSTRAINTS statement.
>
>
> "Statement" means "interactive command" in that context --- in other
> words, the constraints won't be checked until after control returns from
> your function.
Thanks very much, that clarifies this behaviour of PostgreSQL -> it's
definitively a feature, not a bug :-)
--
Dr. Christian Rank
Rechenzentrum Universität Passau
Innstr. 33
D-94032 Passau
GERMANY
Tel.: 0851/509-1838
Fax: 0851/509-1802
PGP public key see http://www.rz.uni-passau.de/mitarbeiter/rank