Show TOC

Using TransientProviders Derived from Classic InfoSetsLocate this document in the navigation structure

Use

Before continuing with this topic, please refer to TransientProvider Derived from a Classic InfoSet.

There may be instances in which the volume of data in an InfoSet becomes so large as to be no longer suitable for direct access. At the same time, it may not warrant a complete, warehoused BI approach.

You can access replicated data using the same TransientProvider as mentioned in relation to direct access above. This provides fast access to replicated data without the overhead of a complete BI system. A replication API sends data to a BIA index, from where it is read by the TransientProvider.

Note

The TransientProvider can access data directly, or it can access replicated data, but not both simultaneously. If you choose to replicate data and access it through the index, there will be a latency corresponding to the time gap between replication and access.



Figure 1: Direct Access and Replicated Access Using a TransientProvider

Partitioning the BIA Index

Partitions allow you to avoid unnecessary replication of data to the BIA index. A partition in the index must be based on a stable criterion, and should reflect a useful division, keeping apart a data slice that is active from another that no longer receives new data, for example. Partitions can be treated independently, allowing you to delete data that is no longer relevant for BI queries without deleting the whole index, and allowing you to update a partition that has new or changed data while leaving an unchanged partition alone.

Note that partitions must be specified as single values. If you choose a partition criterion with a character-like data type, the BI Query can only consider the BIA index if you use the same single value restrictions in the query as you have used for replicating. For numeric fields like fiscal year, an interval selection (range) can be considered.

Replicating to the BIA Index

After choosing to execute data replication, you are given the choice of whether or not to Delete Before Loading.

If the indicator is set, the BIA index is dropped and all data records currently replicated are deleted.

If the indicator is unset, data that is currently replicated to the BIA index is kept. A new data slice complementary to the one currently replicated is loaded in addition.

If your BIA index is partitioned, and you want to extend the current selection condition, you have to select all current partitions. The global selection condition must not vary among different partitions.

Note

If you choose the option not to delete data prior to loading (indicator is unset), only a selection complementary to the current selection condition is loaded.

For more information, see the online help (F1) for the indicator.

More Information

For more information about partitioning and replicating the BIA index, see the online help for transaction BS_ANLY_OP_BIA. On the SAP Easy Access screen, choose Cross-Application Components -> Processes and Tools for Enterprise Applications -> Analytics Infrastructure -> BIA for Classic InfoSets.