Why Loading CSV seeds in dbt? - Purpose & Use Cases

The Big Idea

What if you could turn messy CSV files into clean database tables with just one command?

The Scenario

Imagine you have a list of product details saved in a CSV file. You want to use this data in your database for analysis. Without a simple tool, you might try copying and pasting rows one by one or writing complex scripts to load the data.

The Problem

Manually entering data or writing custom scripts is slow and prone to mistakes. You might miss rows, mix up columns, or spend hours fixing errors. This wastes time and causes frustration.

The Solution

Loading CSV seeds in dbt lets you quickly and reliably bring CSV data into your database as tables. It automates the process, ensuring accuracy and saving you from tedious manual work.

Before vs After

✗ Before

INSERT INTO products VALUES ('123', 'Shoe', 50);
INSERT INTO products VALUES ('124', 'Hat', 20);

✓ After

dbt seed --select products

What It Enables

It makes adding reference data fast and error-free, so you can focus on analyzing and building insights.

Real Life Example

A marketing team loads a CSV of campaign codes and descriptions into their analytics database to join with sales data and measure campaign success.

Key Takeaways

Manual data entry is slow and error-prone.

Loading CSV seeds automates and simplifies data import.

This speeds up workflows and improves data accuracy.

Practice

1. What is the main purpose of loading CSV seeds in dbt?

easy

A. To load small, fixed reference data as tables in the database

B. To run complex SQL transformations on large datasets

C. To create temporary views for data exploration

D. To export data from the database to CSV files

Solution

1. Step 1: Understand the role of seeds in dbt

   Seeds are used to load CSV files as tables in the database, mainly for small, fixed reference data.

2. Step 2: Compare options with seed purpose

   Options B, C, and D describe other dbt or database functions, not seed loading.

3. Final Answer:

   To load small, fixed reference data as tables in the database -> Option A

4. Quick Check:

   Seeds = small fixed reference data [OK]

Hint: Seeds load small fixed data as tables [OK]

Common Mistakes:

• Thinking seeds are for large datasets
• Confusing seeds with models or views
• Assuming seeds export data instead of loading

2. Which folder should you place your CSV files in to load them as seeds in dbt?

easy

A. data

B. models

C. macros

D. snapshots

Solution

1. Step 1: Recall the folder structure for dbt seeds

   CSV files for seeds must be placed in the 'data' folder inside the dbt project.

2. Step 2: Eliminate other folders

   'models' is for SQL models, 'macros' for reusable code, 'snapshots' for snapshot data, so they are incorrect for seeds.

3. Final Answer:

   data -> Option A

4. Quick Check:

   Seeds folder = data [OK]

Hint: Put CSVs in 'data' folder for seeds [OK]

Common Mistakes:

• Placing CSVs in 'models' folder
• Confusing 'macros' with data storage
• Using 'snapshots' folder for seeds

3. Given the following dbt command run in a project with a CSV file named countries.csv in the data folder, what will happen?

dbt seed

medium

A. Nothing will happen unless you specify the table name

B. The CSV file will be deleted from the data folder

C. The CSV file will be loaded as a table named 'countries' in the database

D. The CSV file will be converted to a model SQL file

Solution

1. Step 1: Understand the effect of dbt seed

   This command loads all CSV files in the 'data' folder as tables in the database, using the CSV filename as the table name.

2. Step 2: Apply to the given CSV file

   The file 'countries.csv' will be loaded as a table named 'countries'. No extra arguments are needed.

3. Final Answer:

   The CSV file will be loaded as a table named 'countries' in the database -> Option C

4. Quick Check:

   dbt seed loads CSVs as tables named after files [OK]

Hint: dbt seed loads CSVs as tables named by file [OK]

Common Mistakes:

• Thinking dbt seed deletes files
• Expecting dbt seed to convert CSV to SQL
• Believing table name must be specified manually

4. You ran dbt seed but the table did not appear in your database. Which of the following is the most likely cause?

medium

A. The CSV file is empty

B. The CSV file has a .txt extension instead of .csv

C. You forgot to run dbt run after dbt seed

D. The CSV file is not placed inside the 'data' folder

Solution

1. Step 1: Check the seed loading requirements

   dbt only loads CSV files placed inside the 'data' folder when running dbt seed.

2. Step 2: Analyze the options

   If the CSV is outside 'data', dbt seed ignores it. A .txt file won't be loaded. Running dbt run is unrelated to seeds. An empty CSV still creates an empty table.

3. Final Answer:

   The CSV file is not placed inside the 'data' folder -> Option D

4. Quick Check:

   CSV must be in 'data' folder for seed [OK]

Hint: CSV must be in 'data' folder to load [OK]

Common Mistakes:

• Assuming dbt run loads seeds
• Ignoring file extension importance
• Thinking empty CSV prevents table creation

5. You have a CSV file products.csv with columns id, name, and price. You want to load it as a seed and then create a model that selects only products with price > 100. Which steps should you follow?

hard

A. Place products.csv in 'data', run dbt run, then create a model filtering price > 100

B. Place products.csv in 'data', run dbt seed, then create a model SQL selecting from the seed table filtering price > 100

C. Place products.csv in 'models', run dbt run, then filter price in the CSV file

D. Place products.csv in 'snapshots', run dbt seed, then create a model selecting all products

Solution

1. Step 1: Load CSV as seed

   Place the CSV in the 'data' folder and run dbt seed to load it as a table.

2. Step 2: Create a model filtering data

   Create a model SQL file that selects from the seed table and filters products where price > 100.

3. Step 3: Understand why other options fail

   Place products.csv in 'models', run dbt run, then filter price in the CSV file places CSV in wrong folder and filters CSV manually. Place products.csv in 'data', run dbt run, then create a model filtering price > 100 misses running dbt seed. Place products.csv in 'snapshots', run dbt seed, then create a model selecting all products uses wrong folder and does not filter.

4. Final Answer:

   Place products.csv in 'data', run dbt seed, then create a model SQL selecting from the seed table filtering price > 100 -> Option B

5. Quick Check:

   Seed CSV in 'data' + dbt seed + model filter = correct [OK]

Hint: Seed CSV in 'data', run dbt seed, then model filter [OK]

Common Mistakes:

• Placing CSV in wrong folder
• Skipping dbt seed command
• Filtering data outside SQL model