You can store data clusters in special databases in the ABAP Dictionary. These are called ABAP cluster databases, and have a prescribed structure:

This method allows you to store complex data objects with deep structures in a single step, without having to adjust them to conform to the flat structure of a relational database. Your data objects are then available systemwide to every user. To read these objects from the database successfully, you must know their data types.

You can use cluster databases to store the results of analyses of data from the relational database. For example, if you want to create a list of your customers with the highest revenue, or an address list from the personnel data of all of your branches, you can write ABAP programs to generate the list and store it as a data cluster. To update the data cluster, you can schedule the program to run periodically as a background job. You can then write other programs that read from the data cluster and work with the results. This method can considerable reduce the response time of your system, since it means that you do not have to access the distributed data in the relational database tables each time you want to look at your list.

Data clusters are specific to ABAP. Although it is possible to read a cluster database using SQL statements, only ABAP can interpret the structure of the data cluster.

To save data objects in a cluster database, use the statement EXPORT TO DATABASE.

To create a directory of data clusters and read data objects from cluster databases, use the statement IMPORT FROM DATABASE.

To delete a data cluster from a cluster database, use the statement DELETE FROM DATABASE.

For further information about how to access cluster databases using Open SQL, refer to: