Hi,

While adding support for Ingres in my database abstraction library
http://www.jooq.org, I have encountered E_OP039F_BOOLFACT as
documented here:
http://docs.ingres.com/ingres/9.2/in...rrors-from-opf

This error should be reported, so here I am. Here's how to reproduce
it with Ingres Visual SQL 10.1.0 (E_SC0129 Ingres Release II 10.1.0
(a64.win/118). I can provide more Server Info upon request.

Create this database schema:
http://jooq.svn.sourceforge.net/view...e=text%2Fplain

Fill in this data:
http://jooq.svn.sourceforge.net/view...e=text%2Fplain

Now this query fails:
select * from t_author
join t_book on ((t_author.id = t_book.author_id
and t_book.language_id in (select id from
t_language where (upper(cd) in ('DE', 'EN'))))
or exists (select 1 from t_author where 1 = 0))
order by t_book.id asc

Note: The problem is in the "OR EXISTS" clause that is used inside a
"JOIN..ON" condition.
Note also: The above query is just an integration test query and has
no semantic meaning to me.

Cheers
Lukas

On May 31, 5:46*pm, Lukas Eder <lukas.e... (AT) gmail (DOT) com> wrote:
The problem is as the message says - a consistency error in the parse
tree. In this case, you have the option of using traditional syntax -
making it "t_author, t_book" and turning the "on" clause into a
"where" clause. I would expect that to fix it. But it is a bug as
consistency errors shouldn't happen.

On 1 Jun., 15:22, DougI <dougink... (AT) gmail (DOT) com> wrote:
That's what I thought. I don't really need this query syntax right
now, I just wanted to report the bug. Should I also report it
elsewhere, other than on this news group?

http://bugs.ingres.com ?

HTH
--
Jean-Pierre Zuate
La Fage Conseil
01 47 96 30 69
jean-pierre.zuate (AT) lafageconseil (DOT) fr
http://lafageconseil.fr/


2011/6/3 Lukas Eder <lukas.eder (AT) gmail (DOT) com>

On Jun 6, 12:18*pm, "Jean-Pierre Zuate, La Fage Conseil" <jean-
pierre.zu... (AT) lafageconseil (DOT) fr> wrote:
Sounds reasonable, thanks!

On Tuesday, May 31, 2011 10:46:05 PM UTC+1, Lukas Eder wrote:
Hello people.

I came across a simular issue on one of my Ingres databases running 9.2.1 patch 13982 in a linux redhat environment.

After alot of searching and chucking everything but the kitchen sink at it I discovered the issue was with cache_dynamic in the cofig settings. If this is turned on on this verion then Ingres seems to not like binded integer variables being used in an join.

Hope this helps

Ish