Version 2021

Improving the performance of large Intelligent Cubes: Partitioning

If your Intelligent Cubes contain large quantities of data, you can improve their performance by dividing the datasets into multiple segments, called partitions. The partitions are processed simultaneously, distributed across the processor cores of your Intelligence Server.

To partition an Intelligent Cube

  1. In Developer, log in to your project and navigate to the Intelligent Cube.
  2. Right-click the Intelligent Cube, and select Edit. The Intelligent Cube Editor opens.
  3. Choose Data > Configure Intelligent Cube. The Intelligent Cube Options dialog box opens.
  4. Select the Data Partition category. The Data Partition options are displayed.
  5. From the Partition Attribute drop-down, select the partition attribute for the Intelligent Cube. For the criteria to determine an appropriate distribution key, see Requirements for the partition attribute.
  6. In the Number of Partitions field, type the number of partitions to create. The number should be less than or equal to the number of processor cores on your Intelligence Server machine.
  7. If your data warehouse supports parallel SQL queries, enable the Fetch data slices in parallel from the warehouse check box.
  8. Click OK.