In memory database warehouse

• US 9,734,221 B2
• Filed: 09/12/2013
• Issued: 08/15/2017
• Est. Priority Date: 09/12/2013
• Status: Active Grant

• Current Assignee: SAP SE
• Original Assignee: SAP SE
• Inventors: Sarferaz, Siar
• Primary Examiner(s): Obisesan, Augustine K
• Application Number:
  US14/025,574
• Publication Number:
  US 20150074037A1
• Time in Patent Office: 1,433 Days
• CPC Class Codes: G06F 16/22 Indexing; Data structures therefor; Storage structures
• CPC Class Codes: G06F 16/252 between a Database Management System and a front-end application

Abstract

An in memory database system includes a random access memory to store an in memory database table, a stored set of defined views over the database tables, a stored set of analytical services adapted to provide transaction services and database warehouse services, an engine to run the analytical services on the database table, and an interface to interact with users and provide access to the engine, analytical services, and views to perform transactions and data warehouse functions on the database table.

Claims

1. An in-memory data warehouse system comprising:

one or more processors; and

a computer-readable medium coupled with the processor, the computer-readable medium comprising instructions stored thereon that are executable by the one or more processors to cause the in-memory data warehouse system to perform operations comprising;

storing a plurality of virtual data model views as part of a virtual data model layer configured to enrich application data with time-dependent data having a validity period;

wherein each virtual data model view of the plurality of virtual data model views comprises a keep history flag as a property of each data model view that is only set if time-dependent data is maintained for the application data associated with the virtual data model view;

receiving an indication that the keep history flag has been activated in a first virtual data model view of the plurality of virtual data model views, wherein the first virtual data model view is based on one or more tables;

based on determining that the keep history flag has been activated in the first virtual data model view, generating a shadow table with a same structure as the first virtual data model view;

determining that updates were made to at least one table upon which the first virtual data model view is based;

determining whether the updates made to the at least one table have an effect on the first virtual data model view; and

updating the shadow table, based on determining that updates made to the at least one table have an effect on the first virtual data model view.

2. The in-memory data warehouse system of claim 1, wherein determining that updates made to the at least one table have an effect on the first virtual data model view comprises determining additional values for an attribute of the first virtual data model view.

3. The in-memory data warehouse system of claim 1, the operations further comprising:

exposing historical information by enhancing the first virtual data model view.

4. The in-memory data warehouse system of claim 1, the operations further comprising:

processing a transaction on an in-memory database wherein the transaction comprises data having a validity time;

updating the first virtual data model view in response to the transaction being processed; and

storing the validity time information in the shadow table to provide historical access to past time dependent data following expiration of the validity time.

5. The in-memory data warehouse system of claim 1, wherein in order to keep history information, the shadow table is flagged as insert-only and stores only a delta of a tuple with a single timestamp.

6. The in-memory data warehouse system of claim 1, wherein in order to keep history information, the shadow table is flagged as insert-only and stores a full tuple including additional information that is used to determine a time span of its validity.

7. The in-memory data warehouse system of claim 1, wherein the shadow table stores validity period information comprising information that indicates a time span for which the corresponding data is valid.

8. The in-memory warehouse system of claim 1, wherein the shadow table comprises a read-optimized main store of data, and a differential buffer, wherein the differential buffer is a write-optimized insert-only structure.

9. The in-memory warehouse system of claim 8, wherein all inserts, updates, and delete operations are performed on the differential buffer.

10. The in-memory data warehouse system of claim 9, wherein the main store is not touched by any data modifying operation.

11. The in-memory data warehouse system of claim 10, wherein the differential buffer and the main store are merged periodically within a merge process.

12. An in-memory data warehouse system executed method comprising:

storing, using one or more processors, a plurality of virtual data model views as part of a virtual data model layer configured to enrich application data with time-dependent data having a validity period;

wherein each virtual data model view of the plurality of virtual data model views comprises a keep history flag as a property of each data model view that is only set if time-dependent data is maintained for the application data associated with the virtual data model view;

receiving an indication that the keep history flag has been activated in a first virtual data model view of the plurality of virtual data model views, wherein the first virtual data model view is based on one or more tables;

based on determining that the keep history flag has been activated in the first virtual data model view, generating a shadow table with a same structure as the first virtual data model view;

determining that updates were made to at least one table upon which the first virtual data model view is based;

determining whether the updates made to the at least one table have an effect on the first virtual data model view; and

updating the shadow table, based on determining that updates made to the at least one table have an effect on the first virtual data model view.

13. The method of claim 12, wherein determining that updates made to the at least one table have an effect on the first virtual data model view comprises determining additional values for an attribute of the first virtual data model view.

14. The method of claim 12, the operations further comprising:

exposing historical information by enhancing the first virtual data model view.

15. The method of claim 12, further comprising:

processing a transaction on an in-memory database wherein the transaction comprises data having a validity time;

updating the first virtual data model view in response to the transaction being processed; and

storing the validity time information in the shadow table to provide historical access to past time dependent data following expiration of the validity time.

16. The method of claim 12, wherein in order to keep history information, the shadow table is flagged as insert-only and stores only a delta of a tuple with a single timestamp.

17. The method of claim 12, wherein in order to keep history information, the shadow table is flagged as insert-only and stores a full tuple including additional information that is used to determine a time span of its validity.

18. The method of claim 12, wherein the shadow table comprises a read-optimized main store of data, and a differential buffer, wherein the differential buffer is a write-optimized insert-only structure.

19. The method of claim 18, wherein all inserts, updates, and delete operations are performed on the differential buffer.

20. A non-transitory computer readable storage device having instructions to cause one or more processors to perform operations comprising:

storing a plurality of virtual data model views as part of a virtual data model layer configured to enrich application data with time-dependent data having a validity period;

wherein each virtual data model view of the plurality of virtual data model views comprises a keep history flag as a property of each data model view that is only set if time-dependent data is maintained for the application data associated with the virtual data model view;

receiving an indication that the keep history flag has been activated in a first virtual data model view of the plurality of virtual data model views, wherein the first virtual data model view is based on one or more tables;

based on determining that the keep history flag has been activated in the first virtual data model view, generating a shadow table with a same structure as the first virtual data model view;

determining that updates were made to at least one table upon which the first virtual data model view is based;

determining whether the updates made to the at least one table have an effect on the first virtual data model view; and

updating the shadow table, based on determining that updates made to the at least one table have an effect on the first virtual data model view.

21. The non-transitory computer readable storage device of claim 20, wherein determining that updates made to the at least one table have an effect on the first virtual data model view comprises determining additional values for an attribute of the first virtual data model view.

22. The non-transitory computer readable storage device of claim 20, the operations further comprising:

exposing historical information by enhancing the first virtual data model view.

23. The non-transitory computer readable storage device of claim 20, the operations further comprising:

processing a transaction on an in-memory database wherein the transaction comprises data having a validity time;

updating the first virtual data model view in response to the transaction being processed; and

storing the validity time information in the shadow table to provide historical access to past time dependent data following expiration of the validity time.