Re: group by and count(*) behaviour in 8.3

>
> > can you send structure and execution plan?
> Thank you for your request, the execution plan is the one from
> "explain" (I think) but what is the "structure plan"?

no, only structure :) table and fields.

> The problema was a bug on my import in new database!
>
> To avoid future error of this type, how can I ask to postgres wath
> column is it using in "natural join"?
>

Don't use natural join. It has some others disadvantagedness. Use
classic JOIN. It is safe.

SELECT FROM tab1 JOIN tab2 ON ....

Regards
Pavel Stehule

> tanks again
> and sorry for my error
> Edoardo
> >
> > Regards
> > Pavel Stehule
> >
> > On 02/01/2008, Edoardo Panfili <edoardo(at)aspix(dot)it> wrote:
> >> I am using this query in 8.3beta4 (compiled from source) in MacOS X 10.5.1
> >>
> >> SELECT webName,count(*) FROM contenitore NATURAL JOIN cartellino WHERE
> >> contenitore.tipo='e' GROUP BY webName;
> >>
> >> this is the result
> >> webName | count
> >> --------------------------------------------------+-------
> >> test palermo | 36679
> >> Herbarium Camerinensis - CAME | 36679
> >> Herbarium Universitatis Aeserniae - IS | 36679
> >> Herbarium Universitatis Civitatis Perusii - PERU | 36679
> >> Herbarium Anconitanum - ANC | 36679
> >> Test database - São Paulo | 36679
> >> Herbarium Universitatis Genuensis - GE | 36679
> >> Herbarium Universitatis Senensis - SIENA | 36679
> >> Segnalazioni Siena | 36679
> >> Herbarium Aquilanum - AQUI | 36679
> >> (10 rows)
> >>
> >> but 36679 is the total number of row of the table.
> >> The same query in 8.1.4 retrieves the aspected result (the number of
> >> elements for each webName).
> >>
> >> Is this a bug or a change in the semantic of SQL?
> >>
> >> thank you
> >> Edoardo
> >>
> >> --
> >> Jabber: edoardopa(at)talk(dot)google(dot)com
> >> tel: 075 9142766
> >>
> >> ---------------------------(end of broadcast)---------------------------
> >> TIP 5: don't forget to increase your free space map settings
> >>
>
>
> --
> Jabber: edoardopa(at)talk(dot)google(dot)com
> tel: 075 9142766
>
> ---------------------------(end of broadcast)---------------------------
> TIP 9: In versions below 8.0, the planner will ignore your desire to
> choose an index scan if your joining column's datatypes do not
> match
>