MicroStrategy ONE

Evaluating your application

Features that make an application suitable for a partitioned dataset

An application that meets at least one of the following criteria may be a suitable candidate for a partitioned dataset:

  • Your documents and dashboards are centrally managed and allow your users to analyze data from different perspectives.
  • All the data that you need for the application can be loaded in a single dataset.
  • All tables that have more than two billion rows of data can be split based on the same attribute.

  • Your base Key Performance Indicators (KPIs) are calculated using basic aggregation functions such as Sum, Average, Minimum, Maximum, Count, and so on. The KPIs can be calculated individually for each partition and then combined.

    Once you have created your dataset, you can create derived metrics that use any of the standard MicroStrategy functions.

    For a full list of the most efficient functions to use in partitioned datasets, see Efficient Functions for Partitioned Datasets.

  • The dataset for your application needs to be incrementally updated on a schedule.
  • The dataset for your application is less than or equal to two terabytes (TB) in size.
  • Your data is structured, and an Extract, Transform, Load (ETL) process has been performed on it.

Features that make an application unsuitable for partitioned datasets

An application that meets at least one of the following criteria may not be suitable for a partitioned dataset:

  • Your dataset needs to support self-service analyses, where your users can create their own reports, documents, or dashboards.
  • All of the data for the application cannot be loaded in a single dataset.
  • Your application allows users to add or update data in your warehouse by using Transaction Services.
  • All tables that have more than two billion rows of data cannot be partitioned based on the same attribute.

  • The calculations for your KPIs require the entire dataset. For example, KPIs that use functions such as First, Last, Standard Deviation, OLAP functions, and so on require the entire dataset.

    For a full list of the most efficient functions to use in partitioned datasets, see Efficient Functions for Partitioned Datasets.

  • Your data is unstructured, and include data sources other than RDBMS or flat files.
  • Your dataset needs to be updated in real time.