Re: indexes are fucked

Thanks Ragnar. That solved the problem. I still would
like some explanation about this voodoo. Most
importantly, how can I estimage a "good" statistics
number?

thanks,

Eugene

--- Dr NoName <spamacct11@yahoo.com> wrote:

> > the seqscan is cheaper when a large enough
> > proportion
> > (maybe 5%) of rows are retrieved, and indeed the
> > cost
> > is estimated at 39014
>
> very good explanation. thank you.
>
>
> > try to increase statistics for this column:
> >
> > ALTER TABLE render ALTER COLUMN person_id
> >   SET STATISTICS 1000;
> > ANALYZE render;
> >
> > 1000 is the maximum value, and probably overkill,
> > but you might start with that. If this helps, you
> > can
> > try to lower values until you find the lowest one
> > that still suits your data. Usually, 200 is
> enough.
>
>
> ok, we're getting into voodoo territory. What is
> this
> "statistics"? How can I see what the current value
> is?
> How can I estimate what's a "good" value? Where can
> I
> find more information about this?
>
> thanks,
>
> Eugene
>
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam
> protection around
> http://mail.yahoo.com
>
> ---------------------------(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
>




____________________________________________________
Start your day with Yahoo! - make it your home page
http://www.yahoo.com/r/hs