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 -> Fetch accross commit

Fetch accross commit

From: Chuck <chuckh_nospam_at_softhome.net>
Date: 18 Dec 2003 19:26:29 GMT
Message-ID: <Xns945592E7932ECchuckhsofthomenet@130.133.1.4>


Oracle 8.1.7 on Win2k

I have to delete a large number of rows from a very large table. I cannot do the delete as a single transaction as it will generate gigabytes of rollback data and blow out the largest RBS I have. If I delete individual rows inside a cursor loop with a commit every N rows, will that cause unnecesarry performance problems? For example, will the select be executed again after each commit? I remember being told this would happen by someone way back in version 7.1.

The purge will run after hours and there should be no other selects going against the table so I should not create any ora-1555's.

I am not deleting a high enough % of the rows to justify recreating the table with just the rows I want to keep. Plus there are several Fk's involved and I really don't want to have to go through recreating them which I would have to do if I took the "create a new table" route.

-- 
Chuck
Remove "_nospam" to reply by email
Received on Thu Dec 18 2003 - 13:26:29 CST

Original text of this message

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