Documenting models in YAML in dbt - Practice Problems & Coding Challenges

Challenge - 5 Problems

🧠 Conceptual

intermediate

Understanding YAML structure for dbt model documentation

Which YAML structure correctly documents a dbt model named sales_data with a description and a column named order_id that has its own description?

A

models:
  - sales_data:
      description: "Contains sales order records"
      columns:
        - order_id:
            description: "Unique identifier for each order"

B

models:
  - name: sales_data
    description: "Contains sales order records"
    columns:
      - name: order_id
        description: "Unique identifier for each order"

C

models:
  - name: sales_data
    description: "Contains sales order records"
    columns:
      order_id:
        description: "Unique identifier for each order"

D

models:
  sales_data:
    description: "Contains sales order records"
    columns:
      order_id:
        description: "Unique identifier for each order"

💡 Hint

Remember that in dbt YAML, models are listed as a sequence with - name: and columns are a list of dictionaries with - name:.

The correct YAML structure for documenting dbt models uses a list under models where each model is an item with a name key. Columns are also listed as a sequence with each column having a name and description. Option B follows this exact structure.

Option B uses mapping for models and columns incorrectly. Option B nests keys improperly. Option B mixes list and mapping incorrectly for columns.

❓ Predict Output

intermediate

Output of parsing a dbt model YAML snippet

Given this YAML snippet for a dbt model documentation, what is the value of the description for the column customer_id?

models:
  - name: customer_orders
    description: "Orders placed by customers"
    columns:
      - name: order_id
        description: "Order unique ID"
      - name: customer_id
        description: "Unique customer identifier"

A null

B "Order unique ID"

C "Orders placed by customers"

D "Unique customer identifier"

💡 Hint

Look for the column named customer_id and read its description.

The YAML defines two columns under the model customer_orders. The column customer_id has the description "Unique customer identifier". This is the value asked for.

🔧 Debug

advanced

Identify the error in this dbt model YAML documentation

What error will occur when dbt tries to parse this YAML documentation for a model?

models:
  - name: product_sales
    description: "Sales data for products"
    columns:
      name: product_id
      description: "ID of the product"

A KeyError because 'columns' is not a list

B TypeError because description is not a string

C SyntaxError due to missing dash before column name

D No error, YAML is valid

💡 Hint

Check how columns should be listed in dbt YAML documentation.

The columns key must be a list of dictionaries, each with a name and description. Here, columns is a dictionary missing the dash - to indicate a list item. This causes dbt to treat it as a dictionary, leading to a KeyError when it expects a list.

❓ data_output

advanced

Number of documented columns in a dbt model YAML

How many columns are documented in this YAML snippet?

models:
  - name: user_activity
    description: "Tracks user actions"
    columns:
      - name: user_id
        description: "User identifier"
      - name: action_type
        description: "Type of action performed"
      - name: timestamp
        description: "When the action occurred"

A 3

B 2

C 1

D 4

💡 Hint

Count the number of items under the columns list.

There are three columns listed: user_id, action_type, and timestamp. Each has a name and description.

🚀 Application

expert

Visualizing model documentation completeness from YAML

You have a YAML file documenting multiple dbt models. You want to create a bar chart showing the number of columns documented per model. Which Python code snippet using PyYAML and matplotlib correctly extracts the data and plots it?

A

import yaml
import matplotlib.pyplot as plt

with open('models.yml') as f:
    data = yaml.load(f, Loader=yaml.SafeLoader)

models = data['models']
model_names = []
column_counts = []
for model in models:
    model_names.append(model['name'])
    column_counts.append(len(model['columns']))

plt.plot(model_names, column_counts)
plt.xlabel('Model Name')
plt.ylabel('Number of Columns')
plt.title('Columns Documented per Model')
plt.show()

B

import yaml
import matplotlib.pyplot as plt

with open('models.yml') as f:
    data = yaml.safe_load(f)

model_names = []
column_counts = []
for model in data['models']:
    model_names.append(model['name'])
    column_counts.append(len(model['columns']))

plt.barh(model_names, column_counts)
plt.xlabel('Number of Columns')
plt.ylabel('Model Name')
plt.title('Columns Documented per Model')
plt.show()

C

import yaml
import matplotlib.pyplot as plt

with open('models.yml') as f:
    data = yaml.safe_load(f)

models = data['models']
model_names = []
column_counts = []
for model in models:
    model_names.append(model['name'])
    column_counts.append(len(model.get('columns', [])))

plt.bar(model_names, column_counts)
plt.xlabel('Model Name')
plt.ylabel('Number of Columns')
plt.title('Columns Documented per Model')
plt.show()

D

import yaml
import matplotlib.pyplot as plt

with open('models.yml') as f:
    data = yaml.safe_load(f)

models = data['models']
model_names = []
column_counts = []
for model in models:
    model_names.append(model['name'])
    column_counts.append(len(model['columns']))

plt.scatter(model_names, column_counts)
plt.xlabel('Model Name')
plt.ylabel('Number of Columns')
plt.title('Columns Documented per Model')
plt.show()

💡 Hint

Use safe_load for YAML parsing and bar chart for count visualization. Handle missing columns with get().

Option C correctly uses yaml.safe_load to parse the YAML file safely. It iterates over the models, safely getting the columns list with get to avoid errors if missing. It then plots a bar chart with model names on the x-axis and column counts on the y-axis.

Option C uses yaml.load without specifying a loader, which is unsafe and deprecated. It also uses plt.plot which is a line plot, not ideal for categorical data.

Option C uses a horizontal bar chart which is valid but swaps axes labels incorrectly for the prompt.

Option C uses a scatter plot which is less appropriate for this categorical count data.