In these circumstances I use a control record so that I can filter the "hits" on the file in question by user, account, date range, whatever. It may increase the penalty in terms of performance but since I can switch it off when I have figured out what is happening it works for us.
> >>-----Original Message----- > >>From: u2-users-boun...@listserver.u2ug.org [mailto:u2-users- > >>boun...@listserver.u2ug.org] On Behalf Of Robert Porter > >>Sent: Friday, August 20, 2010 5:48 PM > >>To: u2-users@listserver.u2ug.org > >>Subject: Re: [U2] TCL command history > >> > >>This was for diagnosing what happened. Access to the file is > >>generally programmatically controlled. VERY few of us actually have > >>access to edit a record directly. And from what I can find no one > >>edited the record via TCL. The change was fairly benign and easily > >>corrected, but the fact that the handful of people who have such > >>access all said they didn't touch it prompted an investigation. I've > >>been able through other means to pinpoint a 30 minute window when the > >>record got changed. I also know it was not done though our menu > >>system by the logfile that even less know it exists, and only 2 of us > >>even know where it lives. I'm also currently running a search on the > >>equate for the changed field in the record to see if a program that > >>shouldn't be updating it is out there. Next will be looking for a > >>write without using the equate deck. > >> > >>I've done some trials with using a trigger on another file. However, > >>the performance hit was unacceptable. I'm hoping to revisit triggers > >>after a pending OS and UV upgrade. I hadn't heard of the index idea > >>before... that's a pretty cool thought. I may play with that. > >> > >>Thanks all for the input. > >>Rob > >> > >> > >>>>> Wols Lists <antli...@youngman.org.uk> 8/20/2010 3:53 PM >>> > >>On 20/08/10 15:08, Robert Porter wrote: > >>> Are TCL command histories stored in a file by any chance? I'm > >>trying to find how a record got updated, and it doesn't appear to > >>have gone through our standard program which updates the record with > >>a who/when timestamp. > >>You can do it with an i-descriptor, or a trigger ... you can force a > >>file to log that info. So it doesn't matter HOW the file is updated, > >>the > >>info will be recorded. > >> > >>What you want to do is either create an update trigger that adds the > >>date, time and user to the record, or create an i-descriptor that > >>calls > >>SUBR() to write key, date, time and user to another file. If you go > >>down > >>the i-descriptor route, declare that field as an index, and it will > >>be > >>updated every time the record is updated. > >> > >>Cheers, > >>Wol > >>_______________________________________________ > >>U2-Users mailing list > >>U2-Users@listserver.u2ug.org > >>http://listserver.u2ug.org/mailman/listinfo/u2-users > >>_______________________________________________ > >>U2-Users mailing list > >>U2-Users@listserver.u2ug.org > >>http://listserver.u2ug.org/mailman/listinfo/u2-users _______________________________________________ U2-Users mailing list U2-Users@listserver.u2ug.org http://listserver.u2ug.org/mailman/listinfo/u2-users
