Xref: alice comp.databases.oracle.misc:47849 comp.databases.oracle.tools:36277
Path: alice!news-feed.fnsi.net!newsfeed.icl.net!nntp.news.xara.net!xara.net!gxn.net!tank.news.pipex.net!pipex!warm.news.pipex.net!pipex!logica.co.uk!usenet
From: JOC (JOC)
Newsgroups: comp.databases.oracle.misc,comp.databases.oracle.tools
Subject: Re: ORA 1002 - fetch out of sequence problem in reports 2.5
Date: Fri, 26 Nov 1999 10:25:30 GMT
Organization: Logica UK Ltd.
Lines: 46
Message-ID: <383e5f59.4044006@NEWS>
References: <383bd6ed.11239397@NEWS> <81h4ro$1ai$1@nnrp1.deja.com> <383CD138.C4D01CE6@aon.nl>
X-Newsreader: Forte Free Agent 1.11/32.235


>rtproffitt@my-deja.com wrote:
>
<snip>
>>
>> Summary:
>> The cursor is no longer valid.
>>
>> Common Causes:
>> Fetching after cursor returns no data.
>> Example: an infinite loop with no exit:
>>    Open c1;
>>    loop   fetch c1 into x;  End Loop;
>>
>> Second Case:
>> Selecting for Update, and while in the loop updating with a COMMIT.
>>       "commit across fetches in a select for update "
>> The document explains the locking strategy and why this causes
>> the error.
>>     "The solution to this problem is to move the commit
>>     outside the fetch loop or use a simple select rather
>>     than a select for update."
>>
>> Robert Proffitt
>> Beckman Coulter

I knew this - my problem is that there is no infinite loop, and this
is not a select for update.  The query is the master query in a
report.  The returned data is just used to populate the report.  To
make matters worse, this only happens with one specific set of query
parameters.

It's probably insoluable.  I just will have to get their 6000 records
out, in  a readable form, using SQL prob - at least that gives me
bloody sensible error messages.

Thanks for the help though

Jo
>

*************************************************

The usual - anything I say reflects on me, and not upon my current employer.

*************************************************