Re: [PERFORM] Questions on query planner, join types, and work_mem

2010-07-30 Thread Tom Lane
Peter Hussey pe...@labkey.com writes: Using the default of 1MB work_mem, the planner chooses a hash join plan : Hash Left Join (cost=252641.82..11847353.87 rows=971572 width=111) (actual time=124196.670..280461.604 rows=968080 loops=1) ... For the same default 1MB work_mem, a nested loop

[PERFORM] view columns and performance

2010-07-30 Thread Eric Schwarzenbach
I'm wondering whether columns, in the select list of a view that is used in a join, which are not used either as join criteria or in the select list of the overall query, effect the performance of the query. In other words supposed I define a view something like CREATE view MyView AS SELECT

Re: [PERFORM] view columns and performance

2010-07-30 Thread Tom Lane
Eric Schwarzenbach subscri...@blackbrook.org writes: I'm wondering whether columns, in the select list of a view that is used in a join, which are not used either as join criteria or in the select list of the overall query, effect the performance of the query. If the view gets flattened into

Re: [PERFORM] Testing Sandforce SSD

2010-07-30 Thread Karl Denninger
6700tps?! Wow.. Ok, I'm impressed. May wait a bit for prices to come somewhat, but that sounds like two of those are going in one of my production machines (Raid 1, of course) Yeb Havinga wrote: Greg Smith wrote: Greg Smith wrote: Note that not all of the Sandforce drives include a

Re: [PERFORM] Explains of queries to partitioned tables

2010-07-30 Thread Robert Haas
On Mon, Jul 26, 2010 at 4:47 AM, Vlad Arkhipov arhi...@dc.baikal.ru wrote: There is a partitioned table with 2 partitions: drop table if exists p cascade; create table p (  id bigint not null primary key,  ts timestamp); create table p_actual ( check (ts is null) ) inherits (p); create