Re: Window Functions: v07 APIs and buffering strateties

2008/10/28 David Rowley <dgrowley@gmail.com>:
> Hitoshi Harada Wrote:
>
>> As I promised, version 7 of the window functions is now released. At
>> the same time, git repository branch comes back to master.
>>
>> git: http://git.postgresql.org/?p=~davidfetter/window_functions/.git
>> patch: http://umitanuki.net/pgsql/window_functions.patch.20081028.gz
>
> I've been testing for a couple of hours now and just comparing results to
> the results Oracle gives me. Perhaps not the best method but it's faster
> than reading through the spec.
>
> I managed to get a seg-fault with the following query.
>
> select salary,dense_rank() over (order by salary desc)
> from employee
> group by salary;
>
>
> It's the group by that does it.
>
> test=# select salary,dense_rank() over (order by salary desc) from employee
> group by salary;
> server closed the connection unexpectedly
>        This probably means the server terminated abnormally
>        before or while processing the request.
> The connection to the server was lost. Attempting reset: Failed.
>
>
> It seems to be possible to crash not just dense_rank() and rank().
>
> This crashes too.
>
> select relid,AVG(seq_Scan) OVER (ORDER BY relid)
> FROM pg_stat_user_tables
> GROUP BY relid,seq_scan;

sample=# select relid,AVG(seq_Scan) OVER (ORDER BY relid)
sample-# FROM pg_stat_user_tables
sample-# GROUP BY relid,seq_scan;relid |        avg
-------+--------------------16385 | 7.000000000000000016391 | 3.500000000000000016394 | 2.3333333333333333
(3 rows)

Hmm, I tested it on my environment, but doesn't crash. If you have
further information on that, please tell me. And yes, I haven't paid
attention for such cases. I think at least regression test is
necessary to be added.

> Looking forward to seeing windowing functions in 8.4!

Thanks for your testing.

Regards,


-- 
Hitoshi Harada