RE: Sorting Large Amounts of Data- Oracle to SAS, Part Deux
Date: Wed, 12 May 2010 11:56:21 -0400
Message-ID: <D98F41646724244FAC1DC0B4462220250392645A_at_SERVER719C.VIPCALLING.CORP>
If oracle version is prior to 11g, you can use /*+ BYPASS_UJVC */ hint.
From: oracle-l-bounce_at_freelists.org
[mailto:oracle-l-bounce_at_freelists.org] On Behalf Of Kellyn Pedersen
Sent: Wednesday, May 12, 2010 10:33 AM
To: oracle Freelists
Subject: Sorting Large Amounts of Data- Oracle to SAS, Part Deux
OK, so I'm looking for some help with this one again- I've found a compromise that has been quite successful for both groups so far: sorted hash cluster.
I've been able to easily sort 1.4 billion rows to SAS without a blip on Oracle's side- now here's the challenge that I need help with:
Can I do it in parallel?? If you knew my company, you'd understand-
"Let's just grab it all as fast as we can and use every resource on the
server!" :) I can't seem to find any parallel hint to force
parallelism in the explain plans. Everytime I do get parallel, I lose
the hash on the cluster and as it disregards it, also losing the sort.
Here's the very simple, elegant query... :) Select * from affinity_summary_sort;
Here's the table DDL:
create table AFF_SUM_SORT
(
IBID NUMBER,
IBHID NUMBER,
IND_MOD_KEY NUMBER,
RECENCY_KEY NUMBER,
VARIABLE_CD VARCHAR2(20),
VARIABLE_VALUE NUMBER
)
cluster SORTED_AFFSUM_CLSTR (RECENCY_KEY, IBID, IBHID);
Here's my explain plan:
Description Owner Object Cost Rows Bytes Temp I/O Cost CPU Cost SELECT STATEMENT, GOAL = ALL_ROWS 1 1471655172 94185931008 1 14986 SORT GROUP BY NOSORT 1 1471655172 94185931008 1 14986 TABLE ACCESS HASH DMART_SCHMA AFF_SUM_SORT 1 1471655172 94185931008 1 14986
Any ideas? Can this be forced to perform a parallel exec when a sorted hash cluster exists or does this bypass the cluster then?
Thanks for any and all assistance on this one!
Kellyn Pedersen
Sr. Database Administrator
I-Behavior Inc.
http://www.linkedin.com/in/kellynpedersen
www.dbakevlar.blogspot.com <http://www.dbakevlar.blogspot.com/>
"Go away before I replace you with a very small and efficient shell
script..."
-- http://www.freelists.org/webpage/oracle-lReceived on Wed May 12 2010 - 10:56:21 CDT