Ok, so 300 times the statistics target!
Thanks a lot for the exact code and formula!
On Mon, Aug 1, 2011 at 4:49 PM, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Why? This is purely an implementation detail. From the user's
> viewpoint, either the stats are good enough or they're not --- the exact
> number of rows sampled doesn't seem that interesting.
>
> regards, tom lane
Because the manual says that i can influence the number of rows that
are sampled by increasing the statistics target.
But i don't know how sensitive that value is, i want to know what it
is that i am about to do.
btw i don't like that the number of statistics records are coupled
with the number of rows sampled. it's not only confusing, it doesn't
seem optimal to me.
I guess i will look up the paper to find out more. I didn't make an
academic study about it (yet :P ) .
Cheers!
WBL
--
"Patriotism is the conviction that your country is superior to all
others because you were born in it." -- George Bernard Shaw