Using the dbt retry command to re-execute failed Runs: A Hands-On Example

📘 Introduction

In this hands-on dbt tutorial, we’ll walk you step-by-step through how to use the dbt retry command to efficiently re-run only failed models after a broken dbt run. Instead of re-executing your entire DAG, you can save time and resources by retrying only what’s necessary.

📌 This is a must-know topic for the dbt Analytics Engineering Certification Exam, so mastering it now puts you one step closer to passing the exam and leveling up your data engineering skills! 👨‍🎓

✅ Prerequisites

☑️ A dbt project set up
☑️ Existing dbt models defined in your models/ directory

💡 What is the dbt retry command?

The dbt retry command allows you to:

• Re-run only the models that failed in your previous dbt run
• Automatically skip successful models to save time
• Continue building your DAG from the point of failure without starting over

Let’s walk through a real scenario where dbt retry becomes your best friend.

🧪1️⃣ Existing dbt models in your dbt project

We’ve already created some dbt models in our project.

The data linage of our project looks as follows:

🚀2️⃣ Run models

Let’s run our project with:

dbt run

But we encounter an error. The model transformed_course_mapping fails, and because of that, its downstream dependency dim_courses also fails to run.

All other models that do not depend on the broken model execute successfully.

🛠️3️⃣ Fix Bug

Looking at the error, the issue is with a column name. The query used a column names that doesn’t exist. The correct name isname. So let's open the model and rename the column:

SELECT DISTINCT
    code,
    name,
    source_name
FROM {{ ref("cleaned_course_mapping") }}

🔄4️⃣ Use dbt retry

Now instead of re-running the entire DAG, we can run: