Re: [GENERAL] Instances where enable_seqscan = false is good

2007-09-04 Thread Gregory Stark
"Ow Mun Heng" <[EMAIL PROTECTED]> writes: > Have not changed anything in that area. Question is.. Do I need to? or > should I try out something just to see how it is? > (any) Recommendations would be good. Sorry, I don't have all the original plans. Can you post the explain analyze with and with

Re: [GENERAL] Instances where enable_seqscan = false is good

2007-09-03 Thread Ow Mun Heng
On Tue, 2007-09-04 at 05:15 +0100, Gregory Stark wrote: > "Ow Mun Heng" <[EMAIL PROTECTED]> writes: > > > On Mon, 2007-09-03 at 11:31 +0100, Gregory Stark wrote: > >> "Ow Mun Heng" <[EMAIL PROTECTED]> writes: > >> > > >> > How can I persuade PG to use the index w/o resorting to setting seqscan > >

Re: [GENERAL] Instances where enable_seqscan = false is good

2007-09-03 Thread Gregory Stark
"Ow Mun Heng" <[EMAIL PROTECTED]> writes: > On Mon, 2007-09-03 at 11:31 +0100, Gregory Stark wrote: >> "Ow Mun Heng" <[EMAIL PROTECTED]> writes: >> > >> > How can I persuade PG to use the index w/o resorting to setting seqscan >> > = false >> >> The usual knob to fiddle with is random_page_cost.

Re: [GENERAL] Instances where enable_seqscan = false is good

2007-09-03 Thread Ow Mun Heng
On Tue, 2007-09-04 at 10:06 +0800, Ow Mun Heng wrote: > On Mon, 2007-09-03 at 11:31 +0100, Gregory Stark wrote: > > "Ow Mun Heng" <[EMAIL PROTECTED]> writes: > > >-> Bitmap Heap Scan on drv (cost=30.44..4414.39 > > > rows=1291 width=24) (actual time=62.980..142.594 rows=12591 lo

Re: [GENERAL] Instances where enable_seqscan = false is good

2007-09-03 Thread Ow Mun Heng
On Mon, 2007-09-03 at 11:31 +0100, Gregory Stark wrote: > "Ow Mun Heng" <[EMAIL PROTECTED]> writes: > > > > How can I persuade PG to use the index w/o resorting to setting seqscan > > = false > > The usual knob to fiddle with is random_page_cost. If your database fits > mostly in memory you may wa

Re: [GENERAL] Instances where enable_seqscan = false is good

2007-09-03 Thread Gregory Stark
"Ow Mun Heng" <[EMAIL PROTECTED]> writes: > Same query, executed twice, once using seqscan enabled and the other > with it disabled. Difference is nearly night and day. > > > How can I persuade PG to use the index w/o resorting to setting seqscan > = false The usual knob to fiddle with is random_

[GENERAL] Instances where enable_seqscan = false is good

2007-09-03 Thread Ow Mun Heng
Same query, executed twice, once using seqscan enabled and the other with it disabled. Difference is nearly night and day. How can I persuade PG to use the index w/o resorting to setting seqscan = false (actually, I don't know what are the pro or cons - I read posts from the archives far back as