What are partitions in SSAS?
SSAS let us create cube. Cube is a multi-dimensional database. Data will be stored inside cube as dimensions and Fact Tables.
By default in order to store Fact tables (measure group tables) partitions will be created inside cube.
Be default for one fact table one partition will be created. Partition is simply a physical storage unit inside cube.
Note: End user is nothing to do with partition. Her/she never say, I want data from this or that partition.
For him/her it is always going to be single partition.
Can we span single fact table across multiple partition?
Yes, we can. We can create more than one partition for a single fact table.
Is there anything we have to keep in mind while creating multiple partitions for single fact table?
Yes, we have to make sure that each partition contain unique records.
Advantages of having partition:
- Each partition can be stored inside a separate physical drive bringing parallel data access into picture.
- Each partition can have its own aggregation logic.
- Different storage setting can be defined for each partition. Example – one partition supports MOLAP whereas one supports ROLAP
- Each partition can be processed independently
- By keeping historical data and a new data of a particular method in separate partitions processing speed can be improved.
Does partitions store only method group data?
No, in case of aggregating logic is defined, it also stores aggregated data.
Also refer our article on 8 important SQL Server (SSIS, SSAS and SSRS)MSBI Interview questions
For technical trainings on various topics like WCF (Windows Communication Foundation), MVC (Model View Controller), Business Intelligence, Design Patterns, WPF, TFS and Basic fundamentals feel free to contact SukeshMarla@Gmail.com or visit www.sukesh-marla.com
See practical video on explaining SSIS, SSAS and SSRS (part 1) with sample demo: –