Trino query plan analysis
Developer Advocate Lester Martin leads this 3-part webinar series designed to teach you the fundamentals of how Trino massively parallelizes your queries to produce the best possible performance. By the end of the series you will not only be able to conceptually visualize how a query is decomposed into its granular elements, but be able to validate the optimizations you are expecting are occurring and perform corrective actions if needed to improve performance.
You’ll be at one with the Trino query engine itself and the envy of your data team for the knowledge you will learn!
This webinar has passed. See below for session recordings!
Session 1: Parallel processing fundamentals
Trino, which powers Starburst products, and other SQL engines such as Apache Hive & Apache Spark process data in parallel as much as possible. Those who attend this session will:
- Understand how & why datasets are divided into splits
- Be able to explain the decomposition of a query into tasks, stages, and exchanges
- Visualize multi-stage queries that tackle sorting, aggregation, and multiple types of joins
- See the conceptual visualization of a parallelized query aligns with Trino’s generated visual plan at runtime
Session 2: Query plan syntax
This session focuses on what drives the cost-based optimizer (CBO) and how to generate & read a query plan. Participants will learn about:
- Optimizations available with table statistics
- Generating a textual and a visual query plan
- Understand the format of the query plan output to include fragments, exchanges, distribution, estimates, and performance
- Performance with, and without, accurate table statistics
Session 3: Detailed plan analysis
This final session builds upon the first two and focuses completely on reviewing query plans with a focus on spotting potential performance issues and remedies to overcome them. The following optimizations will be demonstrated live:
- Column pruning, predicate pushdown, and partition pruning
- Trino’s approach to tackle sorting and aggregating across two stages
- Join optimizations including broadcasting, reordering, and dynamic filtering
Meet the speaker:
Lester Martin
Developer Advocate at Starburst
