There is an increasing demand, from users across all organizations, for higher BI performance, driven by the rapid growth of data volumes and number of users. In this recipe, we will see how to transform a report into a cube.
In-Memory cubes help reduce data transfer between the RDBMS and the Intelligence Server, thus decreasing network bandwidth utilization.
A faster response helps improve user satisfaction, and ultimately, wide adoption of the platform and productivity.
When we execute an In-Memory cube, no data is displayed. The query is run and the result set is stored in memory ready to be used by reports, documents, or Visual Insight analyses (starting from 9.3.1 analyses are called dashboards). The process of executing a cube is known as publishing. A cube cannot be used if it is not published.
Probably the easiest way to build a cube is converting an existing report. It is a convenient way to see and validate the actual dataset that will be included before publishing it.
We will convert the report named 28 InternetSales Dashboard:
28 InternetSales Cube
.You can right-click on 28 InternetSales Cube in My Reports and open the SQL View. The SELECT
statement sent to the RDBMS is the same as the original report, but the data is stored in memory and on disk for later use (you can find it under C:Program FilesMicroStrategyIntelligence ServerCube<machine name><some more strange chars><Intelligent Cube Instance ID>.cube
).
Cubes can be unloaded from memory to make more space available when they're not used, and reloaded from the context menu by right-clicking on the cube icon in the Administration | System Monitors | Caches | Intelligent Cubes folder. See the screen capture:
To revert the action and load the cube back in memory, right-click again and choose Load in memory: