Top 33 SQL Interview Questions for Data Analysts (Updated for 2024)

Top 33 SQL Interview Questions for Data Analysts (Updated for 2024)

Introduction

Data Analyst SQL interview questions aim to quickly evaluate your ability to extract metrics and work with data using SQL. You’ll usually need to write SQL queries on a whiteboard or in a code editor.

SQL questions cover a range of topics, from asking when to use a GROUP BY statement to challenging you to create a query that supports or disproves a product metrics hypothesis.

The goal remains simple, regardless of the interview format: produce clean SQL code as fast as possible. Data analyst SQL questions fall into three categories:

  • Easy SQL questions - These questions focus on defining SQL features, basic use cases, and differentiating between commands like ´WHERE´ and ´HAVING´. They may also include simple queries.

  • Intermediate SQL questions - These questions require you to write complex queries using joins, sub-queries, self-joins, and window functions. They may also involve They may also include analytics case studies.

  • Hard SQL questions - These questions challenge you to write advanced queries, including the use of indices and complex SQL clauses. They may also include more advanced analytics case studies.

How Is SQL Tested in Data Analyst Interviews?

Writing on a white Board

SQL technical screens are a part of nearly every data analyst interview. In these screens, candidates are asked to answer real-world problems using SQL.

Most commonly, candidates are provided a dataset and asked to write a SQL query to return the desired data.

How do companies test proficiency? There are three main types of SQL interview questions:

  1. Whiteboarding - SQL whiteboard tests are a common part of interviews. In a whiteboard test, you’re required to write SQL queries by hand, which allows companies to assess your understanding of SQL concepts and problem-solving ability.
  2. Coding tests - Many companies ask you to write code and run queries in live interviews. With live coding screens, you can check for syntax errors while you work, and it provides companies a way to see your coding efficiency.
  3. SQL case studies - In case interviews, you’re given a real-world problem and asked to use your SQL skills to solve the problem. These are typically open-ended questions that leave room for analysis and creative problem-solving.

image

Easy SQL Questions for Data Analysts

Easy SQL Interview Questions for Data Analysts

Data analyst interviews and technical screens generally start with beginner SQL questions. There are two main types of easy SQL questions:

  • Basic SQL Queries - You will often be asked basic SQL interview questions that require you to write a query. You might be asked to get the COUNT of a table, make a simple join, or use the HAVING clause.

  • Definitions - Less frequently encountered, these questions ask you to explain technical concepts, compare two or more SQL functions or define how a concept is used.

Here are nine easy SQL questions for data analysts:

1. When would you use the GROUP BY statement?

The GROUP BY clause enables us to compute aggregate values alongside discrete data by collapsing the distinct rows of our database into summary rows that share a particular characteristic.

If you were to attempt to formulate a query selecting a number of rows with a WHERE clause, and then display an aggregate value alongside, you would find that the query would return an error. This is because SQL is not able to display, in a single table, the results of our WHERE query as a list of values conjoined with the aggregate value you are looking for.

2. What are the most common aggregate functions in SQL? What do they do?

An aggregate function performs a calculation on a set of values and returns a single value summarizing the set. The three most common aggregate functions in SQL are: COUNT, SUM, and AVG.

  • COUNT - Returns the number of items of a group.
  • SUM - Returns the sum of ALL or DISTINCT values in an expression.
  • AVG - Returns the average of values in a group (and ignores NULL values).

image

3. What is a unique key in SQL?

In SQL, a unique key is one or more columns or fields that identify a record in a database. Tables can have multiple unique keys, which is a difference between unique keys and primary keys. With unique keys, only one NULL value is accepted for the column and it cannot have duplicate values.

4. What is the difference between UNION and UNION ALL?

UNION and UNION ALL are SQL operators used to concatenate two or more result sets. This allows us to write multiple SELECT statements, retrieve the desired results, then combine them together into a final, unified set.

The main difference between UNION and UNION ALL is that:

  • UNION: only keeps unique records.
  • UNION ALL: keeps all records, including duplicates.

5. What is the difference between a RIGHT JOIN and a LEFT JOIN?

The two most common types of joins in SQL are LEFT and RIGHT. The main difference is that these JOIN operators deal with matched and unmatched rows.

A LEFT JOIN includes all records from the left table and matched rows from the right. A RIGHT JOIN returns all rows from the right table and unmatched rows from the left.

6. What is the difference between a table and a view?

Tables contain data and they are made up of columns and rows. A view is a virtual table, which generally is dependent on data from the table for its display.

One use case for a view can be found if you wanted to look at a subset of data from a table. You could create a view using the SELECT command to query the data.

image

7. What SQL operator is used for pattern matching?

The LIKE operator is used to search for a specific pattern within a column. It is used with a WHERE clause to query specific columns.

LIKE is used with two wildcard characters:

  • % represents a 0, 1 or multiple characters
  • _ (underscore) represents a single character

Here are a few examples:

WHERE Employee Name LIKE '%r' - Finds matches that end with "r"

WHERE Employee NAME LIKE '%gh%'  - Finds matches that include "gh" in any position `

WHERE Employee NAME LIKE '_ch%' - Finds matches with "ch" in the second and third positions
 
WHERE Employee NAME LIKE 'g%r' - Finds matches that start with "g" and end with "r"

8. What command would you use to update data in a table?

You update an existing table with the UPDATE command in SQL. It is used with SET (which includes the updated information) and WHERE to select the specific instance.

Example: In the table ‘Employees’, you want to change the emergency contact, ContactName, for an employee with EmployeeID 3.

UPDATE Employees
SET ContactName = "Bob Smith"
WHERE EmployeeID = 3;

9. Which operator is used to select values within a range? What types of values can be selected?

The BETWEEN operator is used to select values within a range. You can use numbers, texts or dates with BETWEEN.

One important thing to note: the BETWEEN operator includes both the start and end dates.

SELECT EmployeeID
FROM Employees
WHERE EmployeeID BETWEEN 378 AND 492 

10. Write a query that outputs a random manufacturer’s name

Given a table of cars with columns id and make, write a query that outputs a random manufacturer’s name with an equal probability of selecting any name.

Input:

cars table

id make
1 Ford
2 Toyota
3 Toyota
4 Honda
5 Honda
6 Honda

Output:

Column Type
make Text

image

Intermediate SQL Questions for Data Analysts

For mid-level data analyst roles, expect intermediate SQL questions to challenge your knowledge and skill. Medium SQL questions fall into two buckets:

  • Complex SQL queries - Intermediate SQL questions ask you to perform joins, sub-queries, self-joins, and window functions.
  • SQL/Analytics case studies - Many intermediate questions take the form of case studies or ask you to perform analysis on the data you pull from a query.

Here are some medium SQL questions for data analysts:

11. Find the 3 lowest-paid employees that have completed at least 2 projects.

Given tables employees, employee_projects, and projects, find the 3 lowest-paid employees that have completed at least 2 projects.

Note: incomplete projects will have an end date of NULL in the projects table.

12. Given the tables users and rides, write a query to report the distance traveled by each user in descending order.

For this question, you need to accomplish two things: the first is to figure out the total distance traveled for each user_id, and the second is to order from greatest to least each user_id by a calculated distance traveled.

This question has been asked in Uber data analyst interviews.

13. Write a query to find all the users that are currently “Excited” and have never been “Bored” within a campaign.

For this medium SQL problem, assume you work at an advertising firm. You have a table of users’ impressions of ad campaigns over time. Each user_id from these campaigns has an attached impression_id, categorized as either “Excited” or “Bored”. You will need to assess which users are “Excited” by their most recent campaign and have never been “Bored” in any past campaign.

See a full solution to this problem on YouTube:

SQL questions

Note: This is the type of question that might get asked for a marketing analyst job.

image

14. Write a SQL query to select the second-highest salary in the engineering department.

To answer this question, you need the name of the department to be associated with each employee in the employees table to understand which department each employee is a part of.

The “department_id” field in the employees table is associated with the “id” field in the departments table. You can call the “department_id” a foreign key because it is a column that references the primary key of another table, which in this case is the “id” field in the departments table.

Based on this shared field, you can join both tables using INNER JOIN to associate the department name with their employees.

SELECT salary
    FROM employees
    INNER JOIN departments
        ON employees.department_id = departments.id

With the department name in place you can now look at the employees of the Engineering team and sort by salary to find the second highest paid.

15. Given a table of bank transactions, write a query to get the last transaction for each day.

More Context: The table includes the columns: id, transaction_value and created_at (representing the time for each transaction).

Since our goal in this problem is to pull the last transaction from each day, you want to group the transactions by the day they occurred and create a chronological order within each day from which you can retrieve the latest transaction.

To accomplish the task of grouping and order, create a modified version of the bank_transactions table with an added column denoting the chronological ordering of transactions within each day.

To partition by date, you can use an OVER() statement. After partitioning, you should use a descending order so that the first entry in each partition is the last transaction chronologically. Here is how that query can be written:

SELECT (*), ROW_NUMBER() OVER(PARTITION BY DATE(created_at) 
ORDER BY created_at DESC)

AS ordered_time

16. Write a query to debug an error and select the top five most expensive projects by budget-to-employee ratio.

More context: You are given two tables. A projects table and another that maps employees to their projects, called employee_projects. In this question, however, a bug exists that is causing duplicate rows in the employee_projects table.

Example:

Input:

projects table

column type
id INTEGER
title VARCHAR
state_date DATETIME
end_date DATETIME
budget INTEGER

employee_projects table

Column Type
project_id INTEGER
employee_id INTEGER

Output:

Column Type
title VARCHAR
budget_per_employee FLOAT

This is a good example of a logic-based SQL problem. Although there are a few steps to the solution, the actual SQL queries are fairly simple.

HINT: One way to do the debugging is to group by columns project_id simply and employee_id. By grouping by both columns, you are creating a table that provides distinct values on project_id and employee_id, thereby excluding any duplicates.

image

17. You have a table that represents the total number of messages sent between two users by date on Facebook Messenger. Answer these questions:

  • What are some insights that could be derived from this table?
  • What do you think the distribution of the number of conversations created by each user per day looks like?
  • Write a query to get the distribution of the number of conversations created by each user by day in 2020.

This question tests your data sense, as well as your SQL writing skills. It has also appeared in Facebook data analyst interviews.

To answer the first part of the question regarding insights, there are a number of metrics you could evaluate. You can find the total number of messages sent per day, the number of conversations being started, or the average number of messages per conversation. All of these metrics seek to find users’ level of engagement and connectivity.

You can find a full solution on parts one through three in the following YouTube video:

SQL mock

18. Write a SQL query to create a histogram of the number of comments per user in the month of January 2020.

This intermediate SQL question has been asked in Amazon data analyst interviews. Here is a partial answer from Interview Query:

What does a histogram represent, and what kind of story does it tell? In this case, you are interested in using a histogram to represent the distribution of comments each user has made in January 2020. A histogram with bin buckets of size one means that you can avoid the logical overhead of grouping frequencies into specific intervals.

For example, if you want a histogram of size five, you would have to run a SELECT statement like so:

SELECT
    CASE WHEN frequency BETWEEN 0 AND 5 THEN 5 
        WHEN frequency BETWEEN 5 AND 10 THEN 10 etc..

19. Select the largest three departments with ten or more employees and rank them according to the percentage of employees making over 100,000 dollars.

In this problem, you are given two tables: An employees table and a departments table.

Example:

Input:

employees table

Columns Type
id INTEGER
first_name VARCHAR
last_name VARCHAR
salary INTEGER
department_id INTEGER

departments table

Columns Type
id INTEGER
name VARCHAR

Output:

Column Type
percentage_over_100k FLOAT
department_name VARCHAR
number of employees INTEGER

First, break down the question to understand what it’s asking. Specifically, you break the question down into three clauses of conditions:

  • Top three departments by employee count.
  • % of employees making over $\$100,000$ a year.
  • The department must have at least ten employees.

From here, think about how you would associate employees with their department, calculate and display the percentage of employees making over $\$100,000$ a year, and order those results to provide an answer to the original question.

20. Given a table of students and their SAT test scores, write a query to return the two students with the closest test scores by score difference.

Given that this problem is referencing one table with only two columns, you have to self-reference different creations of the same table. It is helpful to think about this problem in the form of two different tables with the same values.

There are two parts to this question:

  • The first part compares each combination of students and their SAT scores.
  • The second part is figuring out which two students’ scores are the closest.

image

21. Write a query to support or disprove the hypothesis: Clickthrough Rate (CTR) is dependent on search rating.

This question provides a table that represents search results on Facebook, including a query, a position, and a human rating.

22. Write a query to get the number of customers that were upsold by purchasing additional products.

For this problem, you are given a table of product purchases. Each row in the table represents an individual product purchase.

Note: If the customer purchased two things on the same day, that does not count as an upsell, as they were purchased within a similar timeframe. We are looking for a customer returning on a different date to purchase a product.

This question is a little tricky because you have to note the dates that each user purchased products. You can’t just group by the user_id and look where the number of products purchased is greater than one because of the upsell condition.

You have to group by both the date field and the user_id to get each transaction broken out by day and user:

SELECT 
    user_id
    , DATE(created_at) AS date
FROM transactions
GROUP BY 1,2

The query above will now give us a user_id and date field for each row. If there exists a duplicate user_id, then you know that the user purchased on multiple days, which satisfies the upsell condition. What comes next?

23. Given the transactions table below, write a query that finds the third purchase of every user.

Note: Sort the results by the user_id in ascending order. If a user purchases two products at the same time, the lower ID field is used to determine which is the first purchase.

Example:

Input:

transactions table

Columns Type
id INTEGER
user_id INTEGER
created_at DATETIME
product_id INTEGER
quantity INTEGER

Output:

Columns Type
user_id INTEGER
created_at DATETIME
product_id INTEGER
quantity INTEGER

Here is a helpful hint for this question: You need an indicator of which purchase was the third by a specific user. Whenever you are thinking of ranking a dataset, it is helpful to immediately think of a specific window function you can use. You need to apply the RANK function to the transactions table. The RANK function is a window function that assigns a rank to each row in the partition of the result set.

SELECT *, RANK() OVER (
    PARTITION BY user_id ORDER BY created_at ASC
        ) AS rank_value
FROM transactions
LIMIT 100

24. Write a query to retrieve the number of users who have posted each of their job listings only once and the number of users who have posted at least one job multiple times.

This is a LinkedIn data analyst interview question. See a full solution to this question on YouTube:

SQL Linkedin Video

25. Write a query to get the top three highest employee salaries for each department.

For this problem, you are given an employees and a departments table.

Note: If the department contains less than three employees, the top two or top one highest salaries should be listed.

Here’s a hint: You need to order the salaries by department. A window function is useful here. Window functions enable calculations within a certain partition of rows. In this case, the RANK() function would be useful. What would you put in the PARTITION BY and ORDER BY clauses?

Your window function can look something like this:

RANK() OVER (PARTITION BY id ORDER BY metric DESC) AS ranks

Note: When you substitute for the actual id and metric fields, make sure the substitutes are relevant to the question asked and aligned with the data provided to you.

26. Write a query to find the number of non-purchased seats for each flight.

In this Robinhood data analyst question, assume you work for a small airline, and you are given three tables: flights, planes, and flight_purchases.

To get the number of unsold seats per flight, you need to get each flight’s total number of seats available and the total seats sold.

You can do an inner join on all 3 tables since the question states that the flight_purchases table does not have entries of flights or seats that do not exist.

To calculate the number of seats per flight, you use GROUP BY on the flight_id together with COUNT() on seat_id to get a count of seats sold. You then do the calculation of the number of total seats on the flight minus the total seats sold to reach how many seats remained unsold.

image

27. Given a transactions table with date timestamps, sample every fourth row ordered by date.

Here’s a hint for this question to get you started: If you are sampling from this table and you want to sample every fourth value specifically, you will probably have to use a window function.

A general rule of thumb to follow is that when a question states or asks for some Nth value (like the third purchase of each customer or the tenth notification sent), then a window function is the best option. Window functions allow us to use the RANK() or ROW_NUMBER() function to provide a numerical index based on a certain ordering.

28. Write a query that returns all of the neighborhoods that have zero users.

More Context: You are given two tables: the first is a users table with demographic information and the neighborhoods they live in, and the second is a neighborhoods table.

This is an intermediate SQL problem that requires you to write a simple query. Our task is to find all the neighborhoods without users. To reframe the task, you need all the neighborhoods that do not have a single user living in them. This means you have to introduce a column in one table but not in the other, such that you can see user counts by neighborhood.

Hard SQL Questions for Data Analysts

data on graph - topic data analyst sql questions

Advanced SQL questions are common for mid- and senior-level data analyst jobs, and they require you to write advanced SQL queries or work through complex logic-based case studies. The two types of questions include:

  • Advanced SQL writing - Writing queries to debug code, using indices to tune SQL queries, and using advanced SQL clauses.
  • Logic-based questions - These questions can be more challenging analytics cases studies or queries that first require you to solve a logic-based problem.

29. An online marketplace company has introduced a new feature that allows potential buyers and sellers to conduct audio chats with each other prior to transacting. Answer the following questions:

  • How would you measure the success of this new feature?
  • Write a query that can represent if the feature is successful or not.

Here’s the full solution to this complex data analytics case study on YouTube:

data analytics case study problem

30. Write a query to get the total three-day rolling average for deposits by day.

For this question, you are given a table of bank transactions with three columns: user_id, a deposit or withdrawal value (determined if the value is positive or negative), and created_at time for each transaction.

Here’s a hint: Usually, if the problem asks to solve for a moving/rolling average, you are provided the dataset in the form of a table with two columns: date and value. This problem is taken one step further as it provides a table of just transactions, with an interest in filtering for deposits (positive value) and removing records representing withdrawals (negative value, e.g. -10).

You also need to know the total deposit amount (sum of all deposits) for each day, as it will factor into calculating the numerator for the rolling three-day average:

rolling three-day avg for day 3 = [day 3 + day 2 + day 1] / 3

31. Write a SQL query that creates a cumulative distribution of the number of comments per user. Assume bin buckets class intervals of one.

To solve this cumulative distribution practice problem, you are given two tables, a users table and a comments table.

frequency cumulative
0 10
1 25
2 27

32. Write a query to display a graph to understand how unsubscribes are affecting login rates over time.

For this question, assume that you work at Twitter. Twitter wants to roll out more push notifications to users because they think users are missing out on good content. Twitter decides to do this in an A/B test. After you release more push notifications, you suddenly see the total number of unsubscribes increase. How would you visually represent this growth in unsubscribes and its effect on login rates?

33. You are given a table of user experiences representing each person’s past employment history. Answer the following:

Write a query to prove or disprove this hypothesis: Data scientists who switch jobs more frequently become managers faster than data scientists that stay at one job for longer.

For this question, you are interested in analyzing the career paths of data scientists. Let’s say that the titles you care about are bucketed into three categories: data scientist, senior data scientist, and data science manager.

Here’s a partial solution to this question:

This question requires a bit of creative problem-solving to understand how you can prove or disprove the hypothesis. The hypothesis is that data scientists who end up switching jobs more often get promoted faster. Therefore, in analyzing this dataset, you can prove this hypothesis by separating the data scientists into specific segments on how often they switch jobs in their careers.

For example, if you looked at the number of job switches for data scientists that have been in their field for five years, you could prove the hypothesis that the number of data science managers increased with the number of times they had switched jobs.

  • Never switched jobs: 10% are managers
  • Switched jobs once: 20% are managers
  • Switched jobs twice: 30% are managers
  • Switched jobs three times: 40% are managers

image

Learn more about Hard SQL Questions

This course is designed to help you learn everything you need to know about working with data, from basic concepts to more advanced techniques

image