Columnar storage brings a completely new set of possibilities to SQL Server. You can get lightning performance in analytical queries right from your data warehouse, without a special analytical database management system. This chapter started by describing features that support analytical queries in SQL Server other than columnar storage. You can use row or page data compression levels, bitmap filtered hash joins, filtered indexes, indexed views, window analytical and aggregate functions, table partitioning, and more. However, columnar storage adds an additional level of compression and performance boost. You learned about the algorithms behind the fantastic compression delivered by columnar storage. This chapter also included a lot of code, showing you how to create and use the nonclustered and the clustered columnstore indexes, including updating the data, creating constraints, and adding additional B-tree nonclustered indexes.

In the next two chapters, you are going to learn about a completely different way of improving the performance of your databases: memory-optimized tables. In addition, this chapter only started you off with analytics in SQL Server; Chapter 13 and Chapter 14 introduce R, a whole analytical language supported by SQL Server 2016. Chapter 15 introduces Python, another language useful for advanced analytics, for which support was added in SQL Server 2017.