On 26 May 2004 11:04:18 -0700, watuni2000_at_yahoo.co.nz (Tyler Smith
Watu) wrote:
>Ed Stevens <nospam_at_noway.nohow> wrote in message news:<6g99b0tj4343bbqmdla8a09rorr70ijvif_at_4ax.com>...
>> On 26 May 2004 05:24:59 -0700, watuni2000_at_yahoo.co.nz (Tyler Smith
>> Watu) wrote:
>>
>> >HP Unix 11.0
>> >Oracle 8.1.7.4
>> >I have a procedure which runs OK sometimes and errs out at other
>> >times with an ORA-1843. This would have been expected to fail
>> >consistently if there was a date mismatch not only occasionally .The
>> >procedure was created with no errors.
>> >
>> >The procedure in question is called via a shell script (if that helps
>> >any)
>> >
>> >Your input will be greatly appreciated?
>> >
>> >TY
>> >
>> >
>> >
>> >if the code will be of any help the heart of the code is here
>> >
>> >
>> >select (sysdate-to_date(t.start_time,'MM/DD/YY:HH24:MI:SS'))
>> >,s.sid,s.serial#,
>> >replace(substr(s.client_info,0,8),',')
>> > Name from v$transaction t,v$session s
>> > where s.taddr=t.addr and to_date(t.start_time,
>> >'DD/MM/YY:HH24:MI:SS') < (SYSDATE-(1/(96)))
>> > and substr(s.client_info,0,2) not in ('PS')
>>
>>
>> Do some occurances of t.start_time have nulls? And sometimes these
>> are picked up by your WHERE clause? Just an off-the-cuff guess.
>
>
>In response to D Rolfe ------
>I am not sure if data should be at a transient stage at anytime.
>When you examine the ACID properties of an RDBMS
>Under consistency there should be no time when
>data is under an interim/transient stage.Data
>will be seen as under the initial or final stage.
>
> ED,s suggestion------- ,except I am
> missing something if a transaction starts
>there will be a start time if it does not it
>should not have a start_time of null.
>
>Thanks for taking the time to respond
>
>TY
Your're right -- I hadn't picked up on the table (view) you were
selecting from.
Received on Wed May 26 2004 - 15:54:08 CDT
