Building a DAG of models in dbt

Introduction

A DAG helps organize your data models so they run in the right order. It shows how models depend on each other.

When you have multiple data models that rely on each other.

When you want to automate running models in the correct sequence.

When you want to understand how data flows through your transformations.

When you need to avoid errors from running models too early.

When you want to improve data pipeline efficiency by running only needed models.

Syntax

dbt

-- model_a.sql
select * from source_table

-- model_b.sql
select * from {{ ref('model_a') }} where condition

-- model_c.sql
select * from {{ ref('model_b') }}

Use {{ ref('model_name') }} to link models and create dependencies.

dbt automatically builds the DAG from these references.

Examples

This references model_a so dbt knows model_b depends on it.

dbt

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

This shows a model filtering data from another model it depends on.

dbt

select * from {{ ref('model_b') }} where status = 'active'

This aggregates data from a dependent model.

dbt

select id, count(*) from {{ ref('model_c') }} group by id

Sample Program

This example has three models. model_b depends on model_a, and model_c depends on model_b. dbt runs them in order.

dbt

/* model_a.sql */
select 1 as id, 'apple' as fruit

/* model_b.sql */
select * from {{ ref('model_a') }} where id = 1

/* model_c.sql */
select fruit, count(*) as count from {{ ref('model_b') }} group by fruit

Important Notes

Always use ref() to create dependencies; do not hardcode table names.

The DAG helps dbt run models in the right order automatically.

You can visualize the DAG using dbt docs generate and dbt docs serve.

Summary

A DAG shows how models depend on each other in dbt.

Use ref() to link models and build the DAG.

dbt uses the DAG to run models in the correct order.

Practice

1.

What does a DAG represent in dbt?

easy

A. The configuration settings for dbt profiles

B. The syntax rules for writing SQL queries

C. The order in which models depend on each other

D. The list of all tables in the database

Solution

1. Step 1: Understand what DAG means in dbt context

   A DAG (Directed Acyclic Graph) shows how models are connected by dependencies.

2. Step 2: Identify the role of DAG in dbt

   dbt uses the DAG to know which models to run first based on dependencies.

3. Final Answer:

   The order in which models depend on each other -> Option C

4. Quick Check:

   DAG = model dependency order [OK]

Hint: DAG shows model dependencies and run order [OK]

Common Mistakes:

• Confusing DAG with SQL syntax
• Thinking DAG lists all tables
• Mixing DAG with dbt config files

2.

Which of the following is the correct way to reference another model in a dbt SQL file?

SELECT * FROM ___

easy

A. ref(model_name)

B. ref('model_name')

C. 'ref(model_name)'

D. ref:"model_name"

Solution

1. Step 1: Recall the syntax for referencing models in dbt

   dbt uses the function ref() with the model name as a string inside parentheses.

2. Step 2: Check each option for correct syntax

   ref('model_name') uses ref('model_name') which is correct; others have syntax errors or wrong quotes.

3. Final Answer:

   ref('model_name') -> Option B

4. Quick Check:

   Use ref('model_name') with quotes [OK]

Hint: Use ref('model_name') with quotes and parentheses [OK]

Common Mistakes:

• Omitting quotes around model name
• Using wrong quote types
• Using colons or other symbols

3.

Given these two models, what is the order dbt will run them?

-- model_a.sql
SELECT * FROM source_table

-- model_b.sql
SELECT * FROM {{ ref('model_a') }}

medium

A. model_a runs first, then model_b

B. model_b runs first, then model_a

C. Both run simultaneously

D. dbt will error due to circular dependency

Solution

1. Step 1: Identify dependencies from ref()

   model_b references model_a using ref(), so model_b depends on model_a.

2. Step 2: Determine run order based on dependencies

   dbt runs model_a first, then model_b to ensure data is ready.

3. Final Answer:

   model_a runs first, then model_b -> Option A

4. Quick Check:

   Dependency order = model_a before model_b [OK]

Hint: Models run in dependency order: referenced first [OK]

Common Mistakes:

• Assuming ref() means reverse dependency
• Thinking models run simultaneously
• Confusing circular dependency errors

4.

What is wrong with this dbt model code snippet?

SELECT * FROM {{ ref(model_a) }}

medium

A. Model name should be uppercase

B. ref() cannot be used inside SELECT

C. Missing FROM keyword

D. Missing quotes around model name in ref()

Solution

1. Step 1: Check syntax of ref() usage

   ref() requires the model name as a string with quotes inside the parentheses.

2. Step 2: Identify the error in the code snippet

   model_a is not quoted, causing a syntax error in dbt compilation.

3. Final Answer:

   Missing quotes around model name in ref() -> Option D

4. Quick Check:

   ref('model_name') needs quotes [OK]

Hint: Always put model names in quotes inside ref() [OK]

Common Mistakes:

• Forgetting quotes around model names
• Thinking ref() can't be in SELECT
• Assuming case sensitivity causes error

5.

You have three models: model_x, model_y, and model_z. model_y references model_x, and model_z references both model_x and model_y. Which of the following is the correct order dbt will run these models?

hard

A. model_x, model_y, model_z

B. model_y, model_x, model_z

C. model_z, model_y, model_x

D. model_x, model_z, model_y

Solution

1. Step 1: Analyze dependencies among models

   model_y depends on model_x; model_z depends on both model_x and model_y.

2. Step 2: Determine run order respecting dependencies

   model_x runs first (no dependencies), then model_y (depends on model_x), then model_z (depends on both).

3. Final Answer:

   model_x, model_y, model_z -> Option A

4. Quick Check:

   Run order respects dependencies [OK]

Hint: Run models so dependencies are built before dependents [OK]

Common Mistakes:

• Running dependent models before their dependencies
• Ignoring multiple dependencies
• Assuming any order works if models reference each other