Re: SQL*loader problem with SEQUENCE
From: Dave Mausner <dma956_at_lulu.acns.nwu.edu>
Date: 1995/11/06
Message-ID: <47lgvf$n1m_at_news.acns.nwu.edu>#1/1
>I'm using SQLLOAD to append a large table. The key (UNIQUE INDEX) to this
>table is generated using the SEQUENCE(COUNT) command in the ctl.
>
>However, on completion of a load I am getting 'Index in Direct Load
>State' messages because of a constraint violation - and if I check I
>can find a duplicate key.
>
>As SQL*load is generating the number for me, why is it creating
>duplicates?
Date: 1995/11/06
Message-ID: <47lgvf$n1m_at_news.acns.nwu.edu>#1/1
In article <515310448wnr_at_istel-is.demon.co.uk>, Andy Hardy <AndyHard_at_istel-is.demon.co.uk> wrote:
>I'm using SQLLOAD to append a large table. The key (UNIQUE INDEX) to this
>table is generated using the SEQUENCE(COUNT) command in the ctl.
>
>However, on completion of a load I am getting 'Index in Direct Load
>State' messages because of a constraint violation - and if I check I
>can find a duplicate key.
>
>As SQL*load is generating the number for me, why is it creating
>duplicates?
this can happen in parallel direct mode because each instance of the load generates its sequence independently but not uniquely because when they start, the record count of the original data appears the same to each instance.
you can load direct, but avoid parallel when using sequences based on count or max value.
-- daveReceived on Mon Nov 06 1995 - 00:00:00 CET