Re: simulating a select returning a million rows?
From: BChase <bsc7080xxmqc_at_myoracleportal.com>
Date: Sat, 13 Jun 2009 07:52:09 -0400
Message-ID: <uk4735l7eujkvk9riisu69b3fu2b9sadvo_at_4ax.com>
One way I read was that could also alter the statistics temporarily to trick the optimizer into thinking there were a million rows in a table that does not. It alters the explain plain accordingly and can alter performance as well.
Date: Sat, 13 Jun 2009 07:52:09 -0400
Message-ID: <uk4735l7eujkvk9riisu69b3fu2b9sadvo_at_4ax.com>
One way I read was that could also alter the statistics temporarily to trick the optimizer into thinking there were a million rows in a table that does not. It alters the explain plain accordingly and can alter performance as well.
On Sat, 13 Jun 2009 11:59:25 +0200, Robert Klemme <shortcutter_at_googlemail.com> wrote:
>On 07.06.2009 02:03, mh_at_pixar.com wrote:
>> Peter Nilsson <airia_at_acay.com.au> wrote:
>>> Malcolm Dew-Jones wrote:
>>>> select rownum from dual connect by rownum <= 1000000
>>> More common is...
>>>
>>> select level from dual connect by level <= 1000000
>>
>> Both of these are working perfectly for me... thanks to
>> both of you!
>
>But keep in mind that if you are investigating execution plans the mere
>fact that you do not query a real table might dramatically change the
>plan in question probably making the exercise useless.
>
>Kind regards
>
> robert
Received on Sat Jun 13 2009 - 06:52:09 CDT