Beginner's guide to Flink SQL query optimization by Muhammet Orazov

00:00 Intro 00:10 About Muhammet Orazov 00:49 Introduction 01:38 Flink SQL optimizations 13:28 Examples 16:42 Fixing a bug 25:54 Conclusion 26:24 Q&A Apache Flink's SQL layer provides a rich, cost-based optimizer that can dramatically improve query performance, but its inner workings often feel complex to the new learners. In this session, I will try to demystify the Flink SQL engine by walking through its core components, from logical plan generation, to optimizer and to physical execution strategies. We will discover: Logical vs. physical planning: How Flink transforms your SQL into an optimized execution graph Rule-based rewrites: Common rule based transformations such as filter pushdown, projection pruning, or join reordering Cost model fundamentals: How Flink SQL engine estimates costs and chooses the best (or cheapest cost) plan Join and window optimization: How the engine chooses joins, windowing and watermark strategies under the hood Along the way, we will also explore the built-in tools for inspecting such as the `EXPLAIN` statement. By the end of this talk, you will have insights how the SQL engines transform your queries, understand how Flink SQL engine optimizes and generates execution plans. Whether you are learning Flink SQL or simply curious about how modern query optimizers work, this session will guide you through each stage of the process and deepen your knowledge of streaming SQL internals.