From: | "Markus Wollny" <Markus(dot)Wollny(at)computec(dot)de> |
---|---|
To: | "Andreas Wenk" <a(dot)wenk(at)netzmeister-st-pauli(dot)de>, "PG-General Mailing List" <pgsql-general(at)postgresql(dot)org> |
Subject: | Re: GIN and GiST index - more in depth info |
Date: | 2009-06-29 16:35:45 |
Message-ID: | 28011CD60FB1724DBA4442E38277F6260D218C3D@hermes.computec.de |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-general |
> For sure I had a look to Oleg Bartunov' s and Teodor Sigaev's
> website at http://www.sai.msu.su/~megera/wiki/ but for me
> it's still not clear how to describe the differences between
> the indexes and the usage scenarios when to use GIN or GiST.
As far as I understand it's a matter of usage scenario. GIN is extremely
slow on updates, I seem to remember somewhere that it's actually often
better to simply recreate the complete index than to update it; GiST's
write performance is not half as bad. On the other hand, GIN is much
faster on reads than GiST. If you've got some data that is read-only in
nature, you'll probably fare better with GIN. If you need frequent
updates, GiST ist the better choice. In certain scenarios you would use
partial indexes to have a GiST index on current, still heavily updated
data, and a GIN index on older, "archived" rows which are not updated
any longer.
Kind regards
Markus
Jede Stimme zahlt, jetzt voten fur die besten Games: www.bamaward.de
Computec Media AG
Sitz der Gesellschaft und Registergericht: Furth (HRB 8818)
Vorstandsmitglieder: Albrecht Hengstenberg (Vorsitzender) und Rainer Rosenbusch
Vorsitzender des Aufsichtsrates: Jurg Marquard
Umsatzsteuer-Identifikationsnummer: DE 812 575 276
From | Date | Subject | |
---|---|---|---|
Next Message | Scott Mead | 2009-06-29 16:59:20 | Re: pasting into psql garbles text |
Previous Message | Craig Ringer | 2009-06-29 16:26:24 | Re: Python client + select = locked resources??? |