Re: [HACKERS] A view just stopped working out of the blue...

Don Baccus wrote:

> This has been working for months, just fine. I've been porting over a bunch
> more stuff from Oracle to this Postgres-based system, and bam! Now any
> select from the view dies with:
>
> unknown node tag 600 in apply_RIR_view

Node tag 600 is T_Query.

> I've tried dropping and rebuilding the table and view in a test database
> and the problem remains. I recall running into problems with other
> operations many moons ago, where a particular node type wasn't being
> handled by a particular operator (the ones I'd seen previously were
> fixed by the excellent 6.5.* versions).
>
> Is this a similar case? I may do a little digging myself tonight, but
> thought I'd ask to see if this rings a bell with anyone. It's a bit
> strange because this view's been working great on this table for so
> long. I added a couple of extra columns to the table recently but
> the view worked immediately afterwards. The stuff I've been porting
> creates views willy-nilly and it's almost like there's an interaction
> taking place, but that doesn't seem right.
>
> It fails in the same manner if I simply declare the view as:
>
> create view users_alertable as select * from users;

I assume the column with the IN constraint is either new or
changed recently. Seems the system generates some subselect
for that and the rewriter is unable to handle this case.

I don't have the time to tackle that, just some hint to push
you into the right direction.

Be careful if hacking inside the rewriter, it's a very
sensitive area!

Jan

--

#======================================================================#
# It's easier to get forgiveness for being wrong than for being right. #
# Let's break this rule - forgive me. #
#========================================= wieck(at)debis(dot)com (Jan Wieck) #