Re: reducing random_page_cost from 4 to 2 to force index scan

On Fri, May 13, 2011 at 3:20 PM, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us> wrote:
> Robert Haas <robertmhaas(at)gmail(dot)com> writes:
>> On Tue, Apr 26, 2011 at 9:04 PM, Merlin Moncure <mmoncure(at)gmail(dot)com> wrote:
>>> The very first thing to check is effective_cache_size and to set it to
>>> a reasonable value.
>
>> Actually, effective_cache_size has no impact on costing except when
>> planning a nested loop with inner index scan.  So, a query against a
>> single table can never benefit from changing that setting.
>
> That's flat out wrong.  It does affect the cost estimate for plain
> indexscan (and bitmap indexscan) plans.

<rereads code>

OK, I agree. I obviously misinterpreted this code the last time I read it.

I guess maybe the reason why it didn't matter for the OP is that - if
the size of the index page in pages is smaller than the pro-rated
fraction of effective_cache_size allowed to the index - then the exact
value doesn't affect the answer.

I apparently need to study this code more.

--
Robert Haas
EnterpriseDB: http://www.enterprisedb.com
The Enterprise PostgreSQL Company