Re: LOBs, PCTVERSION, and ORA-22924: snapshot too old

From: Jonathan Lewis <jonathan_at_jlcomp.demon.co.uk>
Date: Sat, 27 Mar 1999 17:33:13 -0000
Message-ID: <922556428.12046.0.nnrp-04.9e984b29@news.demon.co.uk>

I wonder if the note about the error being misleading is wrong (or perhaps just misleading) in this case.

Dan Halbert has a process that is deleting his rows, not changing the LOBS which are part of the row. Perhaps it is the attempt to roll back the row header deletion that is causing the snapshot problem rather than access to the blob per se.

Now I have a choice - do I experiment with blobs or do I have to find out the anomaly with row migration in Partition tables ?

Jonathan Lewis
Yet another Oracle-related web site: www.jlcomp.demon.co.uk

keith boulton wrote in message <36fceb6e.16490872_at_195.147.246.90>...
>On Fri, 26 Mar 1999 02:24:04 GMT, halbert_at_bbn.com (Dan Halbert) wrote:
>
>I know nothing about blobs, but the Oracle documentation talks about
>pctversion being used to retain old versions of lob data and then
>talks about nothing but updates:
>
>It may be ( guessing ) that this mechanism is to provide read
>consistency within a blob rather than deletion of the blob. ( A
>guiding design principle at Oracle seems to be that no new feature
>will work quite as any reasonable person would expect it to - see row
>migration in partitioned tables ).
>
Received on Sat Mar 27 1999 - 11:33:13 CST