Deleting Attributes and Texts for a Characteristic

You are in the Modeling functional area of the Data Warehousing Workbench. In the InfoObject tree, choose Delete Master Data from the context menu of your InfoObject.

When you delete master data, you can choose whether entries in the SID table of the characteristic are to be retained or deleted.

If you delete the SID table entry for a particular characteristic value, the SID value assigned to the characteristic value is lost. If you reload attributes for this characteristic value later, a new SID value has to be created for the characteristic value. This has a negative effect on the runtime required for loading. In some cases, deleting entries from the SID table can also lead to serious data inconsistencies. This occurs if the list of used SID values generated from the where-used list is not comprehensive.

Delete, Retaining SIDs

For the reasons given above, you should choose this option as standard. Even if, for example, you want to make sure that individual attributes of the characteristic that are no longer needed are deleted before you load master data attributes or texts, the option of deleting master data but retaining the entries from the SID table is also absolutely adequate.

Delete with SIDs

Note that deleting entries from the SID table is only necessary, or useful, in exceptional cases. Deleting entries from the SID table does make sense if, for example, the composition of the characteristic key is fundamentally changed and you want to swap a large record of characteristic values with a new record with new key values.

You can choose whether the texts are deleted.

You can select search mode. The search mode has major effects on the where-used list's runtime The following search modes are available:

• O only one usage per value:

  Once the where-used list finds a value, the system no longer searches for this value in the other InfoProviders. This search mode is the default setting. This mode is useful if you simply want to delete all values that are no longer used but do not need to carry out any more complex searches and want to keep the run-time of the where-used list to a minimum.

• P one usage per value per InfoProvider:

  The system searches for each value in every InfoProvider. Each time a hit is found in an InfoProvider, the search stops in that InfoProvider. This search mode is useful if you are interested in the usage: you can find out which InfoProvider you have to delete from before you can delete them all.

• E one usage is enough:

  After a value is found once, the search stops. This setting is useful if you only want to completely delete the attributes and texts.

• A all usages of all values:

  The system searches for each value in every InfoProvider; each usage is counted. This is a complete where-used list. This search time has the longest runtime and should only be used if you are interested in all the usages.

You can choose to execute the deletion operation in the background.

You can simulate master data deletion. Here, the where-used list is executed without deleting master data. However, you should note: the list of usages may not be complete as no locks are set during the simulation.

Choose Start. The program checks the entries in the master data table one after the other to see if they are used in other objects. If the master data is used, you can only delete records that are not used or you can display the where-used list.

Once the deletion is complete, a message appears and you can display the log. The log displays the where-used list and the list of all deleted records.