You're reading from Azure Data Engineering Cookbook Get well versed in various data engineering techniques in Azure using this recipe-based guide

Product type Paperback

Published in Sep 2022

Publisher Packt

ISBN-13 9781803246789

Length 608 pages

Edition 2nd Edition

Languages

SQL

Tools

Azure

Concepts

Data Engineering

Authors (3):

Ahmad Osama

Nagaraj Venkatesan

Luca Zanna

View More author details

Table of Contents (16) Chapters

Preface

1. Chapter 1: Creating and Managing Data in Azure Data Lake

2. Chapter 2: Securing and Monitoring Data in Azure Data Lake FREE CHAPTER

3. Chapter 3: Building Data Ingestion Pipelines Using Azure Data Factory

4. Chapter 4: Azure Data Factory Integration Runtime

5. Chapter 5: Configuring and Securing Azure SQL Database

6. Chapter 6: Implementing High Availability and Monitoring in Azure SQL Database

7. Chapter 7: Processing Data Using Azure Databricks

8. Chapter 8: Processing Data Using Azure Synapse Analytics

9. Chapter 9: Transforming Data Using Azure Synapse Dataflows

10. Chapter 10: Building the Serving Layer in Azure Synapse SQL Pool

11. Chapter 11: Monitoring Synapse SQL and Spark Pools

12. Chapter 12: Optimizing and Maintaining Synapse SQL and Spark Pools

13. Chapter 13: Monitoring and Maintaining Azure Data Engineering Pipelines

14. Index

Why subscribe?

15. Other Books You May Enjoy

Monitoring table distribution, data skew, and index health using Synapse DMVs

In distributed databases such as Synapse dedicated SQL pools, the table is distributed across multiple nodes by design. When the rows in the table are not evenly distributed across the nodes, data distribution is said to be skewed. Data skew scenarios can have an impact on query performance. This recipe will provide a script based on Synapse Dynamic Management Views (DMVs) that you can use to monitor table skew.

Tables in a dedicated SQL pool have a column store index created by default. Column store indexes store the rows of the table in columnar format, which is optimized for processing analytics workloads. Each column store index in a table is subdivided into segments. A column store segment can be of three states – Open, Closed, or Compressed. For the column store index to be effective, its segments need to meet the following conditions: