Re: B+ versus hash maps - Mailing list pgsql-general

From surabhi.ahuja
Subject Re: B+ versus hash maps
Date
Msg-id 8626C1B7EB748940BCDD7596134632BE39868E@jal.iiitb.ac.in
Whole thread Raw
In response to delete seems to be getting blocked  ("surabhi.ahuja" <surabhi.ahuja@iiitb.ac.in>)
Responses Re: B+ versus hash maps
List pgsql-general
in that case, should i set
enable_seqscan parameter to off at the time of starting postmaster?
 
because i have seen that even thou the index exists it still goes for seq scan
 
thanks
surabhi


From: Jim Nasby [mailto:jnasby@pervasive.com]
Sent: Fri 6/16/2006 1:49 AM
To: surabhi.ahuja
Cc: Michael Fuhr; pgsql-general@postgresql.org
Subject: Re: [GENERAL] B+ versus hash maps

On Jun 15, 2006, at 8:07 AM, surabhi.ahuja wrote:

> is there any way of specifying wht type of index i want, say hash 
> maps instead of the B+ trees.
> someone told me that in the case where duplicates occur(on the 
> indexed field), hash map are better than B+ trees.

http://www.postgresql.org/docs/8.1/interactive/sql-createindex.html

> and also please tell if i can assume that it will use index only 
> and not go for sequential scan,
> again i was told for that i ll have to set the random page cost 
> parameter to 1.

The database will use whatever it thinks is optimal. Use explain 
analyze to see what it's doing.

Unless your database fits (and stays) entirely in memory, you'll 
probably be pretty unhappy with random_page_cost=1.
--
Jim C. Nasby, Sr. Engineering Consultant      jnasby@pervasive.com
Pervasive Software      http://pervasive.com    work: 512-231-6117
vcard: http://jim.nasby.net/pervasive.vcf       cell: 512-569-9461


pgsql-general by date:

Previous
From: Wayne Conrad
Date:
Subject: Re: table has a many to many relationship with itself ... ?
Next
From: Richard Huxton
Date:
Subject: Re: pgadmin window closes abnormally