| From: | Frank Heikens <frankheikens(at)mac(dot)com> | 
|---|---|
| To: | Tadipathri Raghu <traghu(dot)dba(at)gmail(dot)com> | 
| Cc: | Gary Doades <gpd(at)gpdnet(dot)co(dot)uk>, pgsql-performance(at)postgresql(dot)org, Szymon Guz <mabewlun(at)gmail(dot)com> | 
| Subject: | Re: Optimizer showing wrong rows in plan | 
| Date: | 2010-03-28 09:18:27 | 
| Message-ID: | AC9B4F8F-FFD7-426E-A200-F0A1AB565F29@mac.com | 
| Views: | Whole Thread | Raw Message | Download mbox | Resend email | 
| Thread: | |
| Lists: | pgsql-performance | 
Op 28 mrt 2010, om 11:07 heeft Tadipathri Raghu het volgende geschreven:
> Hi All,
>
> I want to give some more light on this by analysing more like this
>
> 1. In my example I have created a table with one column as  
> INT( which occupies 4 bytes)
> 2. Initially it occupies one page of  space on the file that is (8kb).
>
> So, here is it assuming these many rows may fit in this page.  
> Clarify me on this Please.
See these chapters in the manual: http://www.postgresql.org/docs/8.4/interactive/storage.html
The minimum size of a file depends on the block size, by default 8kb: http://www.postgresql.org/docs/8.4/interactive/install-procedure.html
Regards,
Frank
>
> Regards
> Raghavendra
>
>
> On Sun, Mar 28, 2010 at 2:06 PM, Gary Doades <gpd(at)gpdnet(dot)co(dot)uk> wrote:
> On 28/03/2010 8:33 AM, Tadipathri Raghu wrote:
>>
>> Hi Guz,
>>
>> It is assuming that there are 2400 rows in this table. Probably  
>> you've deleted some rows from the table leaving just one.
>>
>> Frankly speaking its a newly created table without any operation on  
>> it as you have seen the example. Then how come it showing those  
>> many rows where we have only one in it.
>> Thanks if we have proper explination on this..
> It's not *showing* any rows at all, it's *guessing* 2400 rows  
> because you've never analyzed the table. Without any statistics at  
> all, postgres will use some form of in-built guess for a table that  
> produces reasonable plans under average conditions. As you've  
> already seen, once you analyze the table, the guess get's much   
> better and therefore would give you a more appropriate plan.
>
> Regards,
> Gary.
>
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Gary Doades | 2010-03-28 09:27:22 | Re: Optimizer showing wrong rows in plan | 
| Previous Message | Tadipathri Raghu | 2010-03-28 09:07:13 | Re: Optimizer showing wrong rows in plan |