Обсуждение: indexing bug?

Hi.

I've created around 4 indexes on a certain table and the CREATE INDEX
command succeeds. When I check for their existence through pgsql with a \d
'table_name'; they're displayed....

However, when I then try to CLUSTER them, the first cluster command runs,
but *all the other indexes disappear* ! So the rest of the cluster commands
fail. When I now do a \d on the table, I see only one index, the one I made
the first cluster on.

This happens on both 7.1.3 and 7.2b2.

Is this a known bug or am I doing something wrong? Any workarounds?

Thanks in advance.

--Arsalan

-------------------------------------------------------------------
People often hate those things which they do not know, or cannot understand.
--Ali Ibn Abi Talib (AS)

On Wed, Nov 21, 2001 at 12:43:44PM +0530, Arsalan Zaidi wrote:
> Hi.
>
> I've created around 4 indexes on a certain table and the CREATE INDEX
> command succeeds. When I check for their existence through pgsql with a \d
> 'table_name'; they're displayed....
>
> However, when I then try to CLUSTER them, the first cluster command runs,
> but *all the other indexes disappear* ! So the rest of the cluster commands
> fail. When I now do a \d on the table, I see only one index, the one I made
> the first cluster on.

If you look up the documentation for CLUSTER, I do beleive it mentions this.

> This happens on both 7.1.3 and 7.2b2.

That's because it's never been fixed.

> Is this a known bug or am I doing something wrong? Any workarounds?

It's documented (at least it used to be). The workaround would to not use
it.

--
Martijn van Oosterhout <kleptog@svana.org>
http://svana.org/kleptog/
> Magnetism, electricity and motion are like a three-for-two special offer:
> if you have two of them, the third one comes free.

Martijn van Oosterhout <kleptog@svana.org> writes:
>> When I now do a \d on the table, I see only one index, the one I made
>> the first cluster on.

> If you look up the documentation for CLUSTER, I do beleive it mentions this.

CLUSTER has a lot of problems, not only that one :-(.  Rewriting it is
on the TODO list ...

            regards, tom lane