Re: IMMEDIATE constraint enforcement does not comply with the SQL standard

>>>>> "Jim" == Jim Finnerty <jfinnert@amazon.com> writes:

 Jim> The consequences of non-compliance is much more insidious than it
 Jim> at first appears. At stake here is whether PostgreSQL supports
 Jim> non-deferrable constraints at all. The purpose of having
 Jim> non-deferrable IMMEDIATE constraints in the standard is to make
 Jim> query optimizations possible.

Unfortunately the possibility of query optimizations for FKs is already
defeated by the issue of referential actions, where we also violate the
spec.

https://wiki.postgresql.org/wiki/PostgreSQL_vs_SQL_Standard#Trigger_firing_relative_to_referential_constraint_actions

(I just added your issue to that entry, which was already there to
discuss the referential actions vs BEFORE trigger timing.)

-- 
Andrew (irc:RhodiumToad)