dbt-trino

Warnings

• breaking Starting with dbt-core v1.8, users must explicitly install both `dbt-core` and `dbt-trino` (e.g., `pip install dbt-core dbt-trino`). Prior to v1.8, installing the adapter implicitly installed `dbt-core`, but this behavior changed to decouple core and adapter versions.
  Fix: Ensure your installation command is `pip install dbt-core dbt-trino`. If upgrading, verify both packages are installed to compatible versions.
• gotcha Trino's SQL dialect can differ from other data warehouses, requiring query modifications. Common differences include function names (e.g., `IFF()` in Snowflake is `IF()` in Trino, `NVL()` is `COALESCE()`, `DATEADD()` is `DATE_ADD()`) and behavior of implicit type coercion. Direct `PIVOT` functionality is also generally not available in Trino SQL.
  Fix: Consult Trino documentation for SQL function equivalents. Use `dbt_utils.pivot` macro or manual aggregation for pivoting. Be mindful of explicit casting for type comparisons.
• gotcha The `merge` incremental strategy (used for upserts) relies on underlying Trino connector support for `MERGE` statements. Some Trino connectors have limited or no `MERGE` support, which can lead to errors.
  Fix: If `merge` fails, consider using the `delete+insert` incremental strategy instead, or verify your specific Trino connector's `MERGE` capabilities.
• gotcha When using AWS Glue as a metastore with Trino, the `on_table_exists='rename'` configuration for table materialization is not supported and will result in a `NOT_SUPPORTED` error. Glue does not support table renames.
  Fix: Set `on_table_exists='drop'` in your model configuration or `dbt_project.yml` when working with AWS Glue.
• gotcha When using OAuth2 authentication, users have reported `Error: header info didn't have x_token_server` after tokens expire. This indicates that automatic token refresh or re-prompting might not occur as expected, requiring manual intervention.
  Fix: Manually delete the expired token from your `keyring` storage. Ensure `keyring` and the `trino[external-authentication-token-cache]` extra are correctly installed and configured. Monitor for upstream fixes in `dbt-trino` or `trino-python-client`.
• deprecated dbt-core v1.10 introduces new deprecation warnings for custom inputs such as unrecognized resource properties, configurations, top-level keys in YAML files, duplicate YAML keys, and unexpected Jinja blocks. Projects migrating to v1.10 or later may encounter these warnings.
  Fix: Review your `dbt_project.yml` and model configurations for adherence to dbt's expected YAML schema. Use the `--warn-error-options '{"warn": ["Deprecations"]}'` flag to explicitly treat new deprecation warnings as warnings, rather than errors if `--warn-error` is enabled.

Install

• pip install dbt-core dbt-trino Recommended installation for dbt-core v1.8+
• pip install dbt-trino Minimal adapter installation
• pip install 'trino[external-authentication-token-cache]' keyring For OAuth2 with token caching

Imports

• dbt-trino

  N/A (Adapter is configured via profiles.yml)

  dbt adapters like dbt-trino are typically not imported directly in Python code for end-user functionality. Interaction is primarily through the dbt CLI and YAML configuration files like `profiles.yml`.

Quickstart

To quickly get started with dbt-trino, first ensure both `dbt-core` and `dbt-trino` are installed. Then, initialize a dbt project using `dbt init` and configure your `profiles.yml` file to define the connection to your Trino cluster. The example shows a basic configuration with no authentication, which can be extended for LDAP, JWT, or OAuth2. After configuration, you can create a simple SQL model and run it using `dbt run`.

import os

# 1. Install dbt-trino (already covered in install instructions)
# pip install dbt-core dbt-trino

# 2. Initialize a dbt project (if you don't have one)
# dbt init my_trino_project

# 3. Configure your profiles.yml (usually in ~/.dbt/profiles.yml)
# Example profiles.yml content for a basic Trino connection:
# my_trino_project:
#   target: dev
#   outputs:
#     dev:
#       type: trino
#       method: none  # Or 'ldap', 'jwt', 'oauth', etc.
#       user: admin
#       password: ""  # Required if method is ldap/kerberos, otherwise optional
#       catalog: hive_catalog # Your Trino catalog (e.g., hive, iceberg)
#       host: localhost
#       port: 8080
#       schema: default_schema # Your Trino schema (database)
#       threads: 4
#       http_scheme: http # or https

# For OAuth2, ensure 'trino[external-authentication-token-cache]' and 'keyring' are installed
# and set method: oauth in profiles.yml

# 4. Create a simple dbt model (e.g., models/my_model.sql)
# -- models/my_model.sql
# {{ config(materialized='view') }}
# SELECT
#   1 as id,
#   'hello' as message
#

# 5. Run your dbt models (execute in your terminal in the dbt project directory)
# dbt debug --target dev
# dbt run --target dev
# dbt test --target dev