Forecast-model-aware data storage for time series data
A system includes multiple memory modules arranged and configured to store data and at least one processor that is operably coupled to the memory modules. The at least one processor is arranged and configured to select an access pattern of a forecast model, determine a storage layout model based on the identified access pattern of the forecast model, and store values in an order defined by the storage layout model using at least one of the memory modules. The order of the stored values enables sequential access to the stored values for use in the forecast model. Implementations of one or more features of the system may be performed by a computer-implemented method and/or a computer program product.
1. An in-memory database system, the system comprising:
a plurality of memory modules arranged and configured to store data, wherein the plurality of memory modules are non-transitory memory modules; and
at least one processor that is operably coupled to the memory modules and that is arranged and configured to:
identify a combination of two different access patterns of a forecast model, wherein the forecast model comprises time series data, and wherein the two different access patterns comprise two different types of access patterns including a season decomposition access pattern and a series alternation access pattern;
determine a storage layout model based on the identified combination of two different access patterns of the forecast model, wherein the storage layout model combines the season decomposition access pattern and the series alternation access pattern of the two different access patterns; and
store values in an order defined by the storage layout model using at least one of the memory modules, wherein the order of the stored values enables sequential access to the stored values for use in the forecast model.
2. The in-memory database system of claim 1 wherein the at least one processor is arranged and configured to sequentially access the stored values for use in the forecast model.
3. The in-memory database system of claim 1 wherein the combination of two different access patterns includes a main time series of data and the stored values are the main time series data for use in the forecast model.
4. The in-memory database system of claim 1 wherein the combination of two different access patterns includes a main time series of data and at least one additional time series of data and the storage layout model combines the main time series of data and the at least one additional time series of data.
5. The in-memory database system of claim 1 wherein the combination of two different access patterns includes a main time series of data and at least one additional time series of data, wherein the at least one additional time series of data includes time shifted main time series of data and the storage layout model combines the main time series of data and the at least one additional time series of data.
6. The in-memory database system of claim 1 wherein the storage layout model defines a two-dimensional array of the stored values.
7. The in-memory database system of claim 1 wherein the storage layout model defines multiple partitions, wherein each of the partitions includes a two-dimensional array of the stored values.
8. A computer-implemented method for adapting a storage layout model that is performed when instructions are executed that are stored on a non-transitory computer readable storage medium, the method comprising:
identifying a combination of two different access patterns a forecast model, wherein the forecast model comprises time series data, and wherein the two different access patterns comprise two different types of access patterns including a season decomposition access pattern and a series alternation access pattern;
determining a storage layout model based on the identified combination of two different access patterns of the forecast model, wherein the storage layout model combines the season decomposition access pattern and the series alternation access pattern of the two different access patterns; and
storing values in an order defined by the storage layout model in at least one of a plurality of memory modules, wherein the order of the stored values enables sequential access to the stored values for use in the forecast model.
9. The method as in claim 8 wherein the memory modules are part of an in-memory database system.
10. The method as in claim 8 further comprising sequentially accessing the stored values for use in the forecast model.
11. The method as in claim 8 wherein the combination of two different access patterns includes a main time series of data and the stored values are the main time series data for use in the forecast model.
12. The method as in claim 8 wherein the combination of two different access patterns includes a main time series of data and at least one additional time series of data and the storage layout model combines the main time series of data and the at least one additional time series of data.
13. The method as in claim 8 wherein the combination of two different access patterns includes a main time series of data and at least one additional time series of data, wherein the at least one additional time series of data includes time shifted main time series of data and the storage layout model combines the main time series of data and the at least one additional time series of data.
14. The method as in claim 8 wherein the storage layout model defines a two-dimensional array of the stored values.
15. The method as in claim 8 wherein the storage layout model defines multiple partitions, wherein each of the partitions includes a two-dimensional array of the stored values.
16. A computer program product, the computer program product being tangibly embodied on a non-transitory computer-readable storage medium and comprising instructions that, when executed by at least one computing device to, are configured to cause the at least one computing device to:
identify a combination of two different access patterns of a forecast model, wherein the forecast model comprises time series, and wherein the two different access patterns comprise two different types of access patterns including a season decomposition access pattern and a series alternation access pattern;
determine a storage layout model based on the identified combination of two different access patterns of the forecast model, wherein the storage layout model combines the season decomposition access pattern and the series alternation access pattern of the two different access patterns; and
store values in an order defined by the storage layout model in at least one of a plurality of memory modules, wherein the order of the stored values enables sequential access to the stored values for use in the forecast model.
17. The computer program product of claim 16 wherein the combination two different access patterns includes a main time series of data and at least one additional time series of data and the storage layout model combines the main time series of data and the at least one additional time series of data.
18. The computer program product of claim 16 wherein the storage layout model defines a two-dimensional array of the stored values.
19. The computer program product of claim 16 wherein the storage layout model defines multiple partitions, wherein each of the partitions includes a two-dimensional array of the stored values.
20. The in-memory database system of claim 1 , wherein the two different types of access patterns further include an information modelling access pattern.