DOCSEARCHbyAlgolia
Submit Docs

dbt Documentation

website·Indexed 41 days ago

https://docs.getdbt.com/

Documentation for the dbt ecosystem, covering adapter development, naming conventions, and project configuration. Includes guides on using the dbt Wizard CLI Beta, managing .dbtignore files, integrating with Google Sheets via the --scopes flag, and configuring dbt MCP servers with Claude Desktop.

Tokens
548.4K
Snippets
2.2K
Records
4K
Agent score
72%

What's inside dbt

  1. Overview of dbt Charts visualization layer

    dbt Charts is a visualization layer that converts YAML files into interactive dashboards by querying dbt models directly. Dashboards are defined as text and stored in Git. Development is performed locally using the dct command-line tool, which reads board files, executes queries via the connection defined in profiles.yml, and provides a browser-based preview with hot reload.

  2. Overview of MetricFlow in dbt

    MetricFlow is a Python library (compatible with Python 3.8 through 3.12) that allows developers to define and query metrics, dimensions, and dimension values in a dbt project. While it can be used locally in dbt platform or self-hosted dbt, a dbt Starter, Enterprise, or Enterprise+ account is required to use the universal Semantic Layer to query metrics in downstream tools.

  3. Overview of dbt development options

    Depending on your infrastructure and needs, dbt can be developed using several different setups:

    • dbt platform: A fully managed environment providing a browser-based IDE (Studio or Canvas), hosted CI/CD, and documentation. It supports both dbt v1 and v2 engines.
    • Self-hosted: Local development using the dbt VS Code extension or manual installation on Windows, macOS, or Linux. This can be used independently or synced with a dbt platform account for CI/CD and documentation.
    • Local + dbt platform: A hybrid approach using local tools (VS Code extension or dbt platform CLI) connected to a dbt platform account to leverage platform features like Insights and Canvas.
    • dbt Wizard: An AI agent for analytics engineering available in the platform and via CLI. It uses a native metadata engine (lineage, model health, semantic definitions) to build, refactor, and document projects.

  4. Overview of the dbt framework

    The dbt framework allows data teams to transform data collaboratively using software engineering best practices such as version control, modularity, CI/CD, and documentation. It consists of two primary components:

    1. The dbt language: The code written within a project, including SQL select statements, Jinja templating, YAML configurations, and tests.
    2. The dbt engine: The component that compiles the project, executes the transformation graph, and produces metadata.

  5. Overview of dbt Wizard AI agent

    dbt Wizard is an AI agent designed for governed data development. It uses a native metadata engine to understand lineage, model health, tests, contracts, and semantic definitions.

    Key Capabilities:

    • Project Understanding: Context-aware analysis of lineage and runtime context.
    • Impact Awareness: Dependency checking before code changes.
    • Safe Validation: Compiles and builds changes before review.
    • Explore Mode: Allows read-only users to ask questions about production data in plain language.

    Availability:

    • dbt platform: Available in Studio IDE and Wizard home tab (Public preview).
    • Local: Available via Wizard terminal/CLI (Public beta) and Wizard Desktop (Private beta).

  6. Introduction to dbt Unit Tests

    Available from dbt v1.8 (or the 'v1 Latest' release track), unit tests allow developers to validate SQL modeling logic using a small set of static inputs before materializing the full model in production. Unlike standard data tests, which run after a model is built to assess data quality, unit tests enable test-driven development by validating functional code logic independently of the production dataset.

  7. Overview of the dbt-databricks adapter

    The dbt-databricks adapter is the recommended way to connect to Databricks (replacing the deprecated dbt-spark connection). It provides better defaults, such as defaulting to Delta format and using merge for incremental models.

    Feature availability by dbt version:

    • dbt-databricks: Available from dbt 1.0
    • Unity Catalog: Available from dbt 1.1 (supports three-part hierarchies: catalog, schema, model name)
    • Python models: Available from dbt 1.3

  8. Overview of dbt Wizard AI agent

    dbt Wizard is an AI agent designed for governed data development in dbt. It handles the full development lifecycle, including investigation, building, validation, and shipping, by grounding its responses in the project's lineage, tests, contracts, and metric definitions.

    Key capabilities include:

    • Asking project-aware questions grounded in project context.
    • Generating documentation, semantic models, tests, and metrics.
    • Building or refactoring models from plain-language prompts.
    • Reviewing file changes as diffs before saving.
    • Running end-to-end tasks in 'agent mode' (automatic or manual approval of changes).
    • Querying production data in 'Explore mode' (available to read-only users without development access).
    • Git task guidance via the wayfinder bar (committing files, creating branches).

  9. Overview of dbt platform development tools

    The dbt platform provides several integrated tools for data development:

    • dbt platform CLI: Used to develop, test, run, and version control dbt projects directly from the local command line.
    • dbt Studio IDE: A browser-based IDE for building, testing, and version controlling dbt models.
    • dbt Canvas: A drag-and-drop visual experience for creating and visualizing dbt models.
    • dbt Wizard: An AI agent with project context that assesses the impact of changes to help ship higher-quality code.
    • Environment Management: Tools to set up separate production and development environments to isolate testing from production data.

  10. Overview of the dbt Studio IDE

    The dbt Studio IDE is a web-based integrated development environment used for building, testing, running, and version-controlling dbt projects. It allows developers to write modular SQL models using the ref() function, compile dbt code into SQL, and execute it directly against a connected database. Key capabilities include generating project documentation, visualizing the Directed Acyclic Graph (DAG), and managing code versions via git directly from the browser.

  11. Configure GCP Private Service Connect (PSC) for self-hosted services

    GCP Private Service Connect (PSC) allows secure, private connectivity between dbt and self-hosted services (VCS, data warehouses, etc.) without exposing them to the public internet. This feature is available for Enterprise+ and Virtual Private tiers.

  12. Manage user identities with SCIM in dbt Enterprise

    SCIM (System for Cross-Domain Identity Management) is available for Enterprise and Enterprise+ plans. It automates user identities and groups to simplify administration and enhance security, supporting real-time user provisioning, deprovisioning, and profile updates, primarily using Okta as the identity provider.