RE: Track all records read from a table

<Paul>
Getting human-readable data out of a journal isn't hard-- and it's the ONLY place on the system where you can guarantee that some human can't change things! Once our auditors understood that, they were content with our using DBU to fix data base issues (with proper permission to use the utility).

The big 'issue' with reading journals is that they have to be translated back to human form. You DSPJRN to an outfile, then CPYF th a specially formatted DB file FMTOPT(*NOCHK) et voila! Data in human readable form, with the date, time, and user of the journalled transaction.

The formatted file consists of: (a) the journal fields as defined in one of IBM's manuals (timestamp/user/etc.) plus (b) all of the fields that define the normal PF on the system, in the same order. Then that CPYF FMTOPT(*NOCHK) moves all the data into the right places (ignore the 'error' that the end-of-record doesn't line up). Then read that file to find out who did what to whom.

And the evidence is as irrefutable as you can get on the iSeries. Just train your auditors!
</Paul>

... you will find a little programm ANZJRN to do it all automatically at my Open Source site: www.bender-dv.de (sorry it's german and the programm is customized for dateformat *YMD - you would have to do an initial CHGJOB, or customize the programm) , just tell it via parameter the filename, lib the name and lib of the outfile and the time intervall you want to have and it will do the rest for you (DSPJRN, creating a suitable outfile ...

D*B