Book Image

SQL Server 2017 Developer's Guide

Book Image

SQL Server 2017 Developer's Guide

Overview of this book

Microsoft SQL Server 2017 is a milestone in Microsoft's data platform timeline, as it brings in the power of R and Python for machine learning and containerization-based deployment on Windows and Linux. This book prepares you for advanced topics by starting with a quick introduction to SQL Server 2017's new features. Then, it introduces you to enhancements in the Transact-SQL language and new database engine capabilities before switching to a different technology: JSON support. You will take a look at the security enhancements and temporal tables. Furthermore, the book focuses on implementing advanced topics, including Query Store, columnstore indexes, and In-Memory OLTP. Toward the end of the book, you'll be introduced to R and how to use the R language with Transact-SQL for data exploration and analysis. You'll also learn to integrate Python code into SQL Server and graph database implementations as well as the deployment options on Linux and SQL Server in containers for development and testing. By the end of this book, you will be armed to design efficient, high-performance database applications without any hassle.
Table of Contents (25 chapters)
Title Page
Copyright and Credits
Dedication
Packt Upsell
Contributors
Preface
Free Chapter
1
Introduction to SQL Server 2017
Index

Query Store architecture


Query Store is integrated with the query processor in the database engine. A simplified Query Store architecture in SQL Server 2017 is shown in the following figure:

Query Store architecture

Query Store actually has three stores:

  • Query and Plan Store: This stores information about executed queries and execution plans used for their execution
  • Runtime Statistics store: This store holds aggregated execution parameters (execution time, logical reads, and so on) for executed queries within a specified time
  • Wait Stats Store: This store persists wait statistics information

All three stores have instances in memory and persisted representation through disk tables. Due to performance reasons, captured info is not immediately written to disk; rather it is written asynchronously. Query Store physically stores this info into the database primary file group.

When a query is submitted to the database engine and Query Store is enabled for the database, during query compilation, Query...