Øystein Grøvlen

Hash join was introduced in MySQL 8.0.18 and was presented for the first time at FOSDEM 2020. Since then, the performance of hash join has been improved, and I will present results that show this. We will also discuss what kind of queries benefit from hash join, what you need to do for hash join to be used, and how to tune your system for optimal hash join performance.

Two MySQL queries that will return the same result, may sometimes have totally different queries plans. This happens because the query optimizer does not realize that the queries are equivalent. In this presentation, we will discuss how we can rewrite queries to help the optimizer find a better query plan. We will show several examples of how we can transform subqueries to make them more efficient, and we will also discuss how we can identify queries that can become faster if a subquery is replaced by window functions. Finally, we will discuss how MySQL 8.0 can do some of these transformations automatically.

In MySQL 8.0, you can create histograms over column values. Histograms will improve the selectivity estimates used by the query optimizer, especially for conditions on columns that are not indexed. This presentation will cover the types of histograms you can create, and discuss best practices for using histograms. The presentation will contain many practical examples of how histograms improve query execution plans.

MySQL Release 5.6 comes with several new query optimizer features that may give improved performance for your queries. This presentation looks at what kind of queries the different improvements apply to, and we will focus on what you can do to get the most out of these new features. The main topics will be improvements to how MySQL execute disk-bound queries and advances in sub-query optimization. We will also present results from test runs that show significant speed-up of queries in MySQL.