Preparing your MicroStrategy environment and dataset

Before you create your partitioned dataset, you need to ensure that your Intelligence Server machine meets the hardware requirements for loading a partitioned dataset into its memory. Additionally, you must select an appropriate partition attribute that maximizes the performance of your reports, documents, and dashboards.

Requirements for your Intelligence Server machine

To determine whether your Intelligence Server can support your in-memory dataset, note the following:

• The number of partitions you divide your dataset into must be less than or equal to the number of processor cores on your Intelligence Server machine.
• The maximum size of a partition is 2 billion rows.
• Ensure that your Intelligence Server machine has enough memory to handle the data.

Requirements for the metrics on your partitioned dataset

Ensure that the metrics that you need to use can be calculated independently for each partition, then combined.

Requirements for the partition attribute

The partition attribute is the attribute that will be used to divide your dataset. The partition attribute must meet the following requirements:

• The data type of the partition attribute must be one of the following:

  • Integer
  • BigDecimal
  • Text
  • Date
• The partition attribute should be present on as many fact tables as possible, especially your largest fact tables. This requirement ensures that the calculations that you need to perform take full advantage of MicroStrategy's parallel processing capabilities.
• The partition attribute is not used to filter your dataset. This ensures that when users analyze data in your application, the maximum number of partitions are involved in the calculations, which leads to faster response times.
• The partition attribute is on any tables that are larger than two billion rows.