Book Image

MySQL 8 Cookbook

By : Karthik Appigatla
Book Image

MySQL 8 Cookbook

By: Karthik Appigatla

Overview of this book

MySQL is one of the most popular and widely used relational databases in the World today. The recently released MySQL 8 version promises to be better and more efficient than ever before. This book contains everything you need to know to be the go-to person in your organization when it comes to MySQL. Starting with a quick installation and configuration of your MySQL instance, the book quickly jumps into the querying aspects of MySQL. It shows you the newest improvements in MySQL 8 and gives you hands-on experience in managing high-transaction and real-time datasets. If you've already worked with MySQL before and are looking to migrate your application to MySQL 8, this book will also show you how to do that. The book also contains recipes on efficient MySQL administration, with tips on effective user management, data recovery, security, database monitoring, performance tuning, troubleshooting, and more. With quick solutions to common and not-so-common problems you might encounter while working with MySQL 8, the book contains practical tips and tricks to give you the edge over others in designing, developing, and administering your database effectively.
Table of Contents (20 chapters)
Title Page
Dedication
Packt Upsell
Contributors
Preface
Index

Archiving tables


Sometimes, you do not want to keep older data and wish to delete it. If you want to delete all the rows which were last accessed over a month ago, if the table is small (<10k rows), you can straight away use the following:

DELETE FROM <TABLE> WHERE last_accessed<DATE_ADD(NOW(), INTERVAL -1 MONTH)

What happens if the table is big? You know InnoDB creates an UNDO log to restore failed transactions. So all the deleted rows are saved in the UNDO log space to be used to restore in case the DELETE statement aborts in between. Unfortunately, if the DELETE statement is aborted in between, InnoDB copies the rows from the UNDO log space to table, which can make the table inaccessible.

To overcome this behavior, you can LIMIT the number of rows deleted and COMMIT the transaction, running the same thing in a loop until all the unwanted rows are deleted.

This is an example pseudo code:

WHILE count<=0:
    DELETE FROM <TABLE> WHERE last_accessed<DATE_ADD(NOW(), INTERVAL...