Are you operating a data pipeline from various databases into a data warehouse? Do you have tools that drop a lot of file data into a data lake? And your analysts and leadership are constantly asking to combine the data in different ways, complaining that everything is too slow. Trino has been solving this problem for over a decade at companies ranging from Silicon Valley to Europe to India and China. Trino now supports a myriad of data sources including all popular relational databases, all table formats used for modern data lakehouses, and numerous document, key-value, and streaming storage systems. You can query them all in Trino and move even more data back and forth with your SQL queries. The fault-tolerant execution mode of Trino ensures your data is always in good hands. With Java 17 and soon Java 21, native, optimized readers and writers for object storage file formats, cost-based query planning and optimization, and massive horizontal scaling capabilities, Trino is up to tackle even the most demanding workloads. And all that goodness can be managed and controlled with a large selection of supported clients including orchestration systems like dbt or Airflow, and UIs like Superset and others.

In this keynote, Martin will talk about all that and other useful features like table functions, table procedures, modern SQL features like MERGE and extensive support for Python clients. All these aspects allow you to replace a number of other tools and platforms with Trino, reduce complexity, and also improve performance for your pipelines and analytics workloads

Technical Level: Technical Practitioner