Re: BUG #5606: DEFERRABLE and DEFERRABLE INITIALLY DEFERRED are the same

On 6 August 2010 16:25, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Frank Heikens <f(dot)heikens(at)anva(dot)nl> writes:
>>> http://wiki.postgresql.org/wiki/What%27s_new_in_PostgreSQL_9.0#DEFERRABLE_UNIQUE_CONSTRAINTS
>
>>> I don't have to ask PostgreSQL to defer, it works in the second test
>>> as well in the third test without any changes. I guess the example in
>>> the wiki isn't correct, right?
>
>> No, not even a little bit :-(.  I hadn't seen that text; it needs to be
>> corrected.  Will hack on it in a moment.
>
> I edited that wiki section, see what you think.
>

Yes, that's much better. I didn't read this page until just now.

I did, however, read the release notes, and I didn't spot a similar error there:

"""
This allows UPDATE tab SET col = col + 1 to work on columns that have
a unique indexes or are marked as primary key, but DEFERRABLE
INITIALLY DEFERRED must be used to mark the constraint as deferred.
"""

Perhaps this should be something like

"""
This allows UPDATE tab SET col = col + 1 to work on columns that have
unique indexes or are marked as primary keys. If the constraint is
specified as DEFERRABLE it will be checked at the end of the statement
rather than as each row is updated. The constraint check may also be
DEFERRED until the end of the current transaction allowing updates to
be spread over multiple SQL commands.
"""

Regards,
Dean