Re: Illegal date format...

Thomas Gaines wrote:

> Eddie, Andrew, Mark, Peter, et al -
>
> Ouch. I was unpleasantly surprised this morning to find an almost
> identical example of what started this entire thread. I'm running
> Oracle 8.1.7.1.1 on WinNT. I'm definitely NOT pulling your leg.
>
> I'm researching a seemingly unrelated performance issue on a box and
> database that I usually don't deal with. When researching statistics,
> I
> found the following when doing this simple query in SQL*Plus:
>
> SQL> select table_name, last_analyzed from dba_tables where owner =
> 'PALEO'
> 2 and last_analyzed is not null;
>
> TABLE_NAME LAST_ANAL
> ------------------------------ ---------
> C14_CARIACO0 00-DECEMB
> MI_URL 28-SEP-02
>
> I've left my nls_date_format alone. If I apply a date mask, this
> is what I get:
>
> SQL> select table_name, to_char(last_analyzed,'mm/dd/yyyy hh24:mi:ss')
> 2 from dba_tables where owner = 'PALEO' and last_analyzed is not
> null;
>
> TABLE_NAME TO_CHAR(LAST_ANALYZ
> ------------------------------ -------------------
> C14_CARIACO0 00/00/0000 00:00:00
> MI_URL 09/28/2002 11:30:05
>
> This is too weird for words, and I'm _this close_ to declaring a
> portion of this
> database corrupt. I'll continue to research this issue from here
> (running DBV, etc),
> and I'll probably have to export this entire PALEO user, wipe it out,
> and reimport
> it.
>
> My users are definitely not using WebMethods as described earlier in
> this thread,
> and I'm unsure of how the data is being created. My users describe a
> situation
> where a fairly complicated query was running in less than a minute
> just last
> week, and now it takes practically a day to complete. Nothing else
> has
> changed.
>
> Happy day.
>
> TG
>
> shuler_at_ntelos.net (Eddie) wrote in message news:<31c10b07.0209171224.7def9f87_at_posting.google.com>...
> > I saw it with my own eyes! Webmethods inserted an illegal date into
> > the Oracle database. When I query the table here is what I get. We are
> > running database version 7.3.4. And yes the eta_date_lfdat is a date
> > field in the databse. Has anyone seen this before? Eddie Shuler
> >
> > 1* select to_char(eta_date_lfdat,'dd-mon-yyyy') from sap_delivery
> > SQL> /
> >
> > TO_CHAR(ETA_DATE_LFDAT,'DD-MON-YYYY')
> > ---------------------------------------------------------------------------
> > 00-december-0000
> > 00-december-0000

How long has it been since someone went to the alert logs and read them?

Daniel Morgan Received on Wed Oct 02 2002 - 10:24:41 CDT