Re: BUG #13453: PostgreSQL 9.5dev pgbench exponential distribution bug? (when threshold is small)

At 2015-06-18 23:32:21, "Alvaro Herrera" <alvherre(at)2ndquadrant(dot)com> wrote:
>Fabien COELHO wrote:
>>
>> Hello,
>>
>> >\setrandom id 1 50000 exponential 0.0000001
>> >insert into test values (:id);
>>
>> Yes, a very small coefficient on a truncated exponential is pretty close to
>> a uniform distribution. I think that this is a mathematical fact, not a bug.
>
>It's a fact of life also ... you don't notice exponentially-growing

>stuff until it's way out of hand :-)
Thanks, great, it's my miss-understand. I use threshold=1,10,100,1000,10000 and test again, add plot pictures.
http://blog.163.com/digoal(at)126/blog/static/163877040201551891133433/
(The closer to 0 the threshold, the flatter (more uniform) the access distribution.)
>
>--
>Álvaro Herrera http://www.2ndQuadrant.com/
>PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services