Special Features of Write-Optimized DataStore Objects

Because of the special structure of write-optimized DataStore objects, they have some special features when you create a data archiving process.

Tab Page: General Settings

Request-based archiving is the default and cannot be changed.

The reasons for this are:

• The contents of write-optimized DataStore objects are managed by requests.
• The request ID is the primary technical key characteristic of this type of DataStore object.
• The SID of the request ID has a character that increases monotonously with time.

Request-based archiving is a special form of time slice archiving:

Only complete requests can be archived. The time slice condition is converted into a request condition. If is not possible to select at least one complete request with the selected time slice condition, no data is archived.

A technical attribute of the load request (Request Creation Date or Request Load Date for example) can also be selected as a time slice characteristic. In this case, the time slice characteristic is not in the semantic key, and the system derives a condition for a monotonously growing partitioning characteristic from the semantic key. Subsequent processing is the same as for standard DataStore objects.

Tab Page: Selection Profile

In addition to the time characteristics, you can also select the characteristics Load Date of Request or Creation Date of Request as the characteristic for time slices.

For DataStore objects with a semantic key, a characteristic of the semantic key can be selected as the primary partitioning characteristic. For DataStore objects without a semantic key, you can only select a characteristic for time slices. You cannot select a primary partitioning characteristic.

No further partitioning characteristics are allowed.

Tab Page: Semantic Group

The semantic group cannot be changed. This is because of the technical key of the DataStore object. The data is grouped in the archive according to the technical key. A data package loaded into the write-optimized DataStore object corresponds to a data object as seen by (ADK) archiving. This ensures that the order the data was loaded in is retained.