From: | Alvaro Herrera <alvherre(at)2ndquadrant(dot)com> |
---|---|
To: | Gavin Flower <GavinFlower(at)archidevsys(dot)co(dot)nz> |
Cc: | Mark Kirkwood <mark(dot)kirkwood(at)catalyst(dot)net(dot)nz>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, pgsql-hackers(at)postgreSQL(dot)org, marko(at)joh(dot)to |
Subject: | Re: Make ANALYZE more selective about what is a "most common value"? |
Date: | 2017-06-11 03:42:35 |
Message-ID: | 20170611034235.v6xfsdrfzgilrsnh@alvherre.pgsql |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Gavin Flower wrote:
> The standard deviation (sd) is proportional to the square root of
> the number in the sample in a Normal Distribution.
>
> In a Normal Distribution, about 2/3 the values will be within plus
> or minus one sd of the mean.
>
> There seems to be an implicit assumption that the distribution of
> values follows the Normal Distribution - has this been verified?
The whole problem here is precisely to determine what is the data
distribution -- one side of it is how to represent it for the planner
(which we do by storing a number of distinct values, a list of MCVs and
their respective frequencies, and a histogram representing values not in
the MCV list); the other side is how to figure out what data to put in
the MCV list and histogram (i.e. what to compute during ANALYZE).
If we knew the distribution was a normal, we wouldn't need any of these
things -- we'd just store the mean and standard deviation.
--
Álvaro Herrera https://www.2ndQuadrant.com/
PostgreSQL Development, 24x7 Support, Remote DBA, Training & Services
From | Date | Subject | |
---|---|---|---|
Next Message | Mark Kirkwood | 2017-06-11 05:11:15 | Re: Make ANALYZE more selective about what is a "most common value"? |
Previous Message | Alvaro Herrera | 2017-06-11 03:28:48 | Re: Re: BUG #14680: startup process on standby encounter a deadlock of TwoPhaseStateLock when redo 2PC xlog |