Tags

Tags in dex allow you to organize, classify, and filter models based on custom labels. They help teams manage large projects by grouping models by purpose, domain, update frequency, team ownership, or deployment behavior.

Tags are especially useful when:

• You want to run only a subset of models in a Flow

• You need to isolate models owned by a specific team or department

• You want to flag models for a specific DAG layer (e.g. gold, finance, staging)

• You need to run models with specific update schedules (e.g. hourly vs. daily)

Tags are simple, flexible, and easy to apply—yet powerful when used with dex’s orchestration, lineage, and filtering features.

How to Add Tags

Tags are defined in your model using the config() block:

{{ config(
    materialized='table',
    tags=['daily', 'finance']
) }}

select *
from {{ ref('orders') }}

You can assign one or more tags to a model. Tags must be provided as a list of strings.

Where Tags Are Used

Tags in dex can be used in:

• Flows: to run only models with a given tag (e.g. all daily jobs or all marketing models)

• Lineage and Explorer: to visually filter and group models by tag

• Build and CI/CD workflows: to include or exclude specific sets of models

• Governance: to apply review policies, ownership boundaries, or SLA labels

Example Use Cases

Example Use Cases

Tag	Purpose
finance	Identify models used in financial reporting
daily	Run this model on a daily schedule
deprecated	Flag a model for future removal
staging	Mark this as a staging-layer model
ml_inputs	Used in machine learning feature pipelines
external	Consumes data from a third-party or external API

You can create your own taxonomy of tags to reflect how your team thinks about and manages data.

Filtering by Tag in Flows

When configuring a Flow, you can choose to run:

• All models

• Specific models

• Models that have a given tag

This is useful for setting up specialized Flows such as:

• Running only hourly jobs

• Rebuilding all gold layer models

• Testing only experimental models in a dev environment

Example Project Snippet

{{ config(
    materialized='incremental',
    tags=['ml_inputs', 'hourly']
) }}

select
  customer_id,
  avg(session_length) as avg_session_time
from {{ ref('web_sessions') }}
group by customer_id

This model is tagged for use in both machine learning pipelines and hourly flows.

Best Practices

• Use lowercase, underscore-separated names (e.g. marketing_reports, not MarketingReports)

• Keep your tag vocabulary consistent across the project

• Use tags to reflect functional or operational intent (e.g. hourly, external, mart)

• Combine tags with documentation for better discoverability

• Regularly clean up unused or outdated tags