Oracle FAQ Your Portal to the Oracle Knowledge Grid
HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US
 

Home -> Community -> Usenet -> c.d.o.server -> Re: a sad story for Oracle DBAs

Re: a sad story for Oracle DBAs

From: Zhencan Fan <zhen_at_wam.umd.edu>
Date: 12 Apr 1998 20:23:20 GMT
Message-ID: <6gr7ro$4qn@dailyplanet.wam.umd.edu>


jared still (jkstill_at_teleport.com) wrote
: Assuming that MTS is not being used, 18 meg for the SGA on a data warehouse
: may not be so far fetched.

: The fact tables will most likely be queried via full table scans. It is
: very possible that much of the memory is being used for hash joins. Did
: your friend mention what the init.ora parameters were set to for hash
: joins? Probably several megabytes.

As much as I would like to agree with you, what you are saying here is false.
1. data warehousing systems tend to have unpredictable space for tables.

   You simply can't afford to allocate hash clusters for these tables. 2. Full table scans usually take longer under hashing.

: 18 meg may also be enough to cache all of th dimension tables.

: Don't jump to conclusions based on the bare minimum of factual information.

: Jared Still
: Certified Oracle DBA and Part Time Perl Evangelist ;)
: Regence Blue Cross Blue Shield of Oregon
: jkstill_at_bcbso.com
: jkstill_at_teleport.com
Received on Sun Apr 12 1998 - 15:23:20 CDT

Original text of this message

HOME | ASK QUESTION | ADD INFO | SEARCH | E-MAIL US