Deleting Master data - can database level tricks work?

[former_member184494]

Very often , we run in master data issues and wonder how our master data can be cleaned.

We have a quick and dirty option : - Drop the data in the tables using SE14.

Should you exercise this option in the interest of time ? If you delete the master data in the SID Tables - then when you reload the master data - there is a possibility of the SIDs changing.

For instance - Customer A might have had SID 1 and now on deletion and reload - Customer A has an SID of 4

What does this impact :

Cubes - the dimension tables / line item facts / aggregates store SIDs - if you delete and reload data - you can possibly compromise this data.

Nav Attributes etc are also similarly affected.

You will have to execute a whole lot of RSRV to correct the SIDs and till then the data is unreadable .

Option 2 : Use the delete master data option :

This is the safest option albeit a very slow one. We once did this in our staging environment and put all loads and reports on hold. The exercise took about 2 days and cleaned our master data completely keeping only the valid ones.

What happens if you started a master data deletion and it does not finish on time ?

You can kill the job . The data loads might still get locked because of the cancelled deletion.

Use SAP Note 936694 - RSDMD138: Master data/text of characteristic already deleted

This will lead to clearing of some control entries in RSMASD which will bring your loads back on schedule.

What if you want to reduce the length of master data / or change compounding ..?

A safer way is to create a new master data object with the new format - and use this new object instead if possible. There are other ways to keep master data consistent and flexible too... Will try and document the same in a separate blog.

Please let me know how you managed master data changes..