Multi-source fan-in patterns in dbt - Practice Problems & Coding Challenges

Challenge - 5 Problems

❓ Predict Output

intermediate

What is the output of this dbt model SQL?

Given two source tables sales and returns, this dbt model combines them using a UNION ALL and calculates net sales. What will be the output of the final query?

dbt

with sales_data as (
  select customer_id, sum(amount) as total_sales
  from {{ source('raw', 'sales') }}
  group by customer_id
),
returns_data as (
  select customer_id, sum(amount) as total_returns
  from {{ source('raw', 'returns') }}
  group by customer_id
),
combined as (
  select customer_id, total_sales, 0 as total_returns from sales_data
  union all
  select customer_id, 0 as total_sales, total_returns from returns_data
)
select customer_id, sum(total_sales) - sum(total_returns) as net_sales
from combined
group by customer_id
order by customer_id

A [{'customer_id': 1, 'net_sales': 100}, {'customer_id': 2, 'net_sales': -50}]

B [{'customer_id': 1, 'net_sales': 150}, {'customer_id': 2, 'net_sales': 0}]

C [{'customer_id': 1, 'net_sales': 100}, {'customer_id': 2, 'net_sales': 50}]

D [{'customer_id': 1, 'net_sales': 0}, {'customer_id': 2, 'net_sales': 50}]

💡 Hint

Think about how UNION ALL combines rows and how sums are aggregated after grouping.

The UNION ALL stacks sales and returns rows. Returns have positive amounts but are subtracted in the final sum, so net sales can be negative if returns exceed sales.

❓ data_output

intermediate

How many rows will the final output have?

Consider a dbt model that merges three source tables orders, shipments, and invoices using UNION ALL without filtering duplicates. If orders has 100 rows, shipments has 80 rows, and invoices has 120 rows, how many rows will the combined output have?

dbt

select * from {{ source('raw', 'orders') }}
union all
select * from {{ source('raw', 'shipments') }}
union all
select * from {{ source('raw', 'invoices') }}

A 120

B 280

C 100

D 300

💡 Hint

UNION ALL stacks all rows without removing duplicates.

UNION ALL combines all rows from the three tables, so total rows equal the sum of rows from each source.

🔧 Debug

advanced

Identify the error in this multi-source fan-in model

This dbt model tries to join two sources and then union with a third source. What error will this code raise?

dbt

with joined as (
  select a.id, a.value, b.status
  from {{ source('raw', 'table_a') }} a
  join {{ source('raw', 'table_b') }} b on a.id = b.id
),
unioned as (
  select * from joined
  union all
  select id, value, status from {{ source('raw', 'table_c') }}
)
select * from unioned

A Runtime error: column mismatch in UNION ALL

B SyntaxError due to missing comma

C No error, query runs successfully

D Runtime error: ambiguous column reference

💡 Hint

Check if the columns selected in both parts of the UNION ALL match exactly.

The first part selects three columns: id, value, status. The second part must select the same three columns in the same order and compatible types. If table_c has different columns or order, UNION ALL fails.

🚀 Application

advanced

Which approach best combines multiple sources with different schemas?

You have three source tables with different columns. You want to create a unified table with columns id, date, and amount. Which dbt SQL snippet correctly uses multi-source fan-in pattern to achieve this?

A

select id, date, amount from {{ source('raw', 'table1') }}
union all
select id, date from {{ source('raw', 'table2') }}
union all
select id, date, amount from {{ source('raw', 'table3') }}

B

select id, date, amount from {{ source('raw', 'table1') }}
union all
select id, date, amount from {{ source('raw', 'table2') }}
union all
select id, date, amount from {{ source('raw', 'table3') }}

C

select id, date, amount from {{ source('raw', 'table1') }}
union
select id, date, amount from {{ source('raw', 'table2') }}
union
select id, date, amount from {{ source('raw', 'table3') }}

D

select id, date, amount from {{ source('raw', 'table1') }}
union all
select id, date, amount, extra_col from {{ source('raw', 'table2') }}
union all
select id, date, amount from {{ source('raw', 'table3') }}

💡 Hint

All SELECT statements in UNION ALL must have the same number of columns and compatible types.

Option B selects the same columns from all tables, ensuring the UNION ALL works correctly. Option B has an extra column in one SELECT, causing error. Option B uses UNION which removes duplicates, not always desired. Option B misses the amount column in one SELECT, causing column mismatch.

🧠 Conceptual

expert

Why use multi-source fan-in patterns in dbt?

Which of the following best explains the main benefit of using multi-source fan-in patterns in dbt projects?

A To enforce strict schema validation on each source table individually

B To improve query performance by splitting data into multiple smaller tables

C To combine data from multiple sources into a single model for easier analysis and reporting

D To automatically generate documentation for all source tables

💡 Hint

Think about why you would want to merge data from different places into one place.

Multi-source fan-in patterns help combine data from different sources into one unified model, making it easier to analyze and report on combined data.