Trino: The Definitive Guide

Chapter 6. Connectors

In Chapter 3, you configured a catalog to use a connector to access a data source in Trino—specifically, the TPC-H benchmark data—and then learned a bit about how to query that data with SQL.

Catalogs are an important aspect of using Trino. They define the connection to the underlying data source and storage system, and use concepts such as connector, schema, and table. These fundamental concepts are described in Chapter 4, and their use with SQL is discussed in more detail in Chapter 8.

A connector translates the query and storage concepts of an underlying data source, such as a relational database management system (RDBMS), object storage, or a key-value store, to the SQL and Trino concepts of tables, columns, rows, and data types. These can be simple SQL-to-SQL translations and mappings but also much more complicated translations from SQL to object storage or NoSQL systems. These can also be user defined.

You can think of a connector the same way you think of a driver for a database. It translates the user input into operations that the underlying database can execute. Every connector implements the Trino service provider interface (SPI). This enables Trino to allow you to use the same SQL tooling to work with whatever underlying data source the connector exposes and makes Trino a SQL-on-Anything system.

Query performance is also influenced by the connector implementation. The most basic connector makes a single connection to the data source and provides ...