Book Image

Learn T-SQL Querying - Second Edition

By : Pedro Lopes, Pam Lahoud
Book Image

Learn T-SQL Querying - Second Edition

By: Pedro Lopes, Pam Lahoud

Overview of this book

Data professionals seeking to excel in Transact-SQL (T-SQL) for Microsoft SQL Server and Azure SQL Database often lack comprehensive resources. This updated second edition of Learn T-SQL Querying focuses on indexing queries and crafting elegant T-SQL code, catering to all data professionals seeking mastery in modern SQL Server versions and Azure SQL Database. Starting with query processing fundamentals, this book lays a solid foundation for writing performant T-SQL queries. You’ll explore the mechanics of the Query Optimizer and Query Execution Plans, learning how to analyze execution plans for insights into current performance and scalability. Through dynamic management views (DMVs) and dynamic management functions (DMFs), you’ll build diagnostic queries. This book thoroughly covers indexing for T-SQL performance and provides insights into SQL Server’s built-in tools for expedited resolution of query performance and scalability issues. Further, hands-on examples will guide you through implementing features such as avoiding UDF pitfalls, understanding predicate SARGability, Query Store, and Query Tuning Assistant. By the end of this book, you‘ll have developed the ability to identify query performance bottlenecks, recognize anti-patterns, and skillfully avoid such pitfalls.
Table of Contents (18 chapters)
1
Part 1: Query Processing Fundamentals
4
Part 2: Dos and Don’ts of T-SQL
9
Part 3: Assembling Our Query Troubleshooting Toolbox

Introducing XEvents

When we connect to the SQL Database Engine and run a query, it fires a series of events – a user logs in, a connection is established, a query begins executing, a plan is found in the cache, a plan is recompiled, and a query completes execution (these are just a few examples). Virtually everything that happens within the Database Engine is an event.

While Dynamic Management Views (DMVs) are powerful tools, they don’t always give a complete picture of what is going on within the engine. Most DMVs provide a snapshot in time, a picture of what is going on the moment they are queried. They may have some history that goes back to the last time the server was restarted, but even then, the information is typically cumulative; they can’t tell us what the server looked like a few minutes before, and they can’t tell us the events that led up to the current state. This is where tracing comes in. Tracing allows us to capture all the occurrences...