Book Image

SQL Server Analysis Services 2012 Cube Development Cookbook

Book Image

SQL Server Analysis Services 2012 Cube Development Cookbook

Overview of this book

Microsoft SQL Server is a relational database management system. As a database, it is a software product whose primary function is to store and retrieve data as requested by other software applications. SQL Server Analysis Services adds OLAP and data mining capabilities for SQL Server databases. OLAP (online analytical processing) is a technique for analyzing business data for effective business intelligence. This practical guide teaches you how to build business intelligence solutions using Microsoft’s core product – SQL Server Analysis Services. The book covers the traditional multi-dimensional model which has been around for over a decade as well as the tabular model introduced with SQL Server 2012. Starting with comparing MultiDimensional and tabular models – discussing the values and limitations of each, you will then cover the essential techniques for building dimensions and cubes. Following on from this, you will be introduced to more advanced topics, such as designing partitions and aggregations, implementing security, and synchronizing databases for solutions serving many users. The book also covers administrative material, such as database backups, server configuration options, and monitoring and tuning performance. We also provide a primer on MultiDimensional eXpressions (MDX) as well as Data Analysis expressions (DAX) languages. This book provides you with data cube development techniques, and also the ongoing monitoring and tuning for Analysis Services.
Table of Contents (19 chapters)
SQL Server Analysis Services 2012 Cube Development Cookbook
Credits
About the Authors
About the Reviewers
www.PacktPub.com
Preface
Index

Partitioning Tabular Models


Partitions in Tabular Models are one of the key features you would use Tabular Models to support. This is the best way to support incremental processing of data in Tabular Models.

Note

Power Pivot models do not support partitions.

Getting ready

In order to implement partitions, you need to have an existing table. You also need to have a partitioning plan. Most often, partitioning makes the most sense when working with data that resembles a fact table with a high row count and lots of measures. We only have one data table in our model, Internet Sales. Order Date, that looks like the best candidate for partitioning. It is unlikely that we will get orders out of sequence that will cause us to reprocess a large number of partitions. Next, we need to determine the scheme we will use. We will focus on a historic partition, a previous six months partition, and a current partition.

How to do it…

Creating partitions in a Tabular Model is fairly straightforward. You will need...