"Bernt Marius Johnsen" writes:
> Description:Query gives different number of rows depending on ORDER
> BY
The attached patch should fix this.
regards, tom lane
Index: src/backend/optimizer/README
===
> "Tom" == Tom Lane writes:
Tom> I'm inclined to think that the best solution is to have
Tom> process_equivalence just reject any clauses that have equal()
Tom> left and right sides, ie, throw them back to be processed as
Tom> ordinary non-equivalence clauses.
>> Hmm. Is it ever possibl
Andrew Gierth writes:
> "Tom" == Tom Lane writes:
> Tom> I'm inclined to think that the best solution is to have
> Tom> process_equivalence just reject any clauses that have equal()
> Tom> left and right sides, ie, throw them back to be processed as
> Tom> ordinary non-equivalence clauses.
>
> "Tom" == Tom Lane writes:
Tom> After digging into it, I find that:
Tom> 1. Without ORDER BY, process_equivalence generates an
Tom> equivalence class that lists k twice. This is pretty bogus but
Tom> it happens to produce the desired results in the example at
Tom> hand. (In some othe
Tom Lane writes:
> Andrew Gierth writes:
>> Notice that the (k = k) qual is being dropped somewhere, which changes
>> the output since that's a disguised not-null condition.
> Huh ... I'm surprised it only does that with the ORDER BY present.
> I suppose it's got something to do with the equival
Andrew Gierth writes:
> Notice that the (k = k) qual is being dropped somewhere, which changes
> the output since that's a disguised not-null condition.
Huh ... I'm surprised it only does that with the ORDER BY present.
I suppose it's got something to do with the equivalence-class machinery.
> Bernt Marius Johnsen wrote:
>> Dump of the database:
To save anyone else the bother, there's a VASTLY simpler testcase for
this one, requiring no tables at all:
test1=# explain select * from (values (1),(null)) v(k) where k = k order by k;
QUERY PLAN
Heikki Linnakangas wrote (2009-09-28 18:10:07):
> Bernt Marius Johnsen wrote:
> > Dump of the database:
Attached.
>
> It looks like the dump got word-wrapped somewhere along the way. Could
> you gzip it and post it again, please?
>
> --
> Heikki Linnakangas
> EnterpriseDB ht
Bernt Marius Johnsen wrote:
> Dump of the database:
It looks like the dump got word-wrapped somewhere along the way. Could
you gzip it and post it again, please?
--
Heikki Linnakangas
EnterpriseDB http://www.enterprisedb.com
--
Sent via pgsql-bugs mailing list (pgsql-bugs@postgresql.org)
