Deleting workitems in a productive system

[Aydin, Kemal (Kemal)** CTR **]

Mike,
 
There is an OSS note to run RSWWCIDE for the "C" type work items. You can search OSS with RSWWCIDE. It's much faster - cause it's a direct delete ...
 
Thanks,
Kemal
 
> -----Original Message-----
>> From: Michael Pokraka [SMTP:workflow at quirky.me.uk]
> Sent: Thursday, February 13, 2003 9:23 AM
> To:   SAP-WUG at MITVMA.MIT.EDU
> Subject:      Deleting workitems in a productive system
>
> Greetings all,
> I have a requirement that's a bit conflicting:
> Client is generating a lot of workflows which have no historical
> significance (They are mostly for processing documents after creation).
>
> Now, we can use archiving, but something irks me about archiving a large
> amount of data (we're talking in Gigabytes/month), the majority of which
> is useless.
> Some alternatives come to mind: deleting, 'split' archiving, or just
> biting the bullet and using standard archiving. Some random thoughts on
> this:
>
>  - Deletion via RSWWHIDE and RSWWWIDE is not recommended by SAP or
>    'The Book' (and also in a recent post by Jocelyn :)
>    Here I would create and thoroughly test a set of variants to be run
>    regularly. However I'm curious as to what (if anything) might not be
>    covered by these jobs.
>  - Archiving these seperately. I'm not exactly an archiving expert and
>    have no idea if it's feasible to split/create variants of the
>    WORKITEM archiving object - one to archive the 'real' workflows and
>    one to archive the 'transient' ones into an electronic dustbin. I
>    don't see any selections based on any specific tasks.
>    If possible, this would be the favoured approach as I believe
>    archiving has a lesser impact on system load.
>  - Biting the bullet and just using normal archives. I have no idea
>    apart from storage requirements if there is any impact on things like
>    retrieval or anything else. Out of principle it's a bit irritating
>    though :-)
>
> Any input would be greatly appreciated.
> Cheers
> Mike