Interviewers don't just want the answer. They want the reasoning, said aloud.
Most SQL interviews are not tests of syntax. You are handed tables you have never seen and asked a question about them, and what is judged is whether you find out what the tables hold before you write anything, whether the query returns the right rows for the right reason, and whether you can say why in plain words. The job that follows asks for the same things, every day.
Explain Your Query answers the questions SQL interviews ask three ways: precisely; aloud, in the words you would use across a table from an interviewer; and on real-shaped tables, with each query and the rows it returned printed together, so that every claim can be checked.
It is written in SQL on DuckDB, and wherever an answer changes between products it says how for PostgreSQL, SQL Server, MySQL, Snowflake and BigQuery, with the pandas equivalent where there is one.
Inside:
- Why SQL is still the language of data, reading tables you have never seen, one standard and many dialects, and how a query runs, from the logical order of its clauses to reading a plan
- SQL against a spreadsheet, against Python and the BI layer, and against non-relational databases, with a rule for choosing in each case
- The interview core: keys, cardinality and duplicates; joins explained by what they do to rows; WHERE and HAVING, the three COUNTs and NULL; CTEs, subqueries and temporary tables; window functions and dates; indexes and slow queries
- A take-home and a live coding test worked from the first clarifying question to the recommendation, and the ordinary week of an analyst
- Ten interview questions with answers shaped to be said aloud, the follow-up each one invites and the common wrong answer, and ten checks to run before you trust a join
The examples use six small tables from HealSync, a fictional hospital network: insurance claims, telemedicine sessions, satisfaction surveys, public-health measurements, research studies and the workforce. They are generated teaching data, awkward in the ways real data is: rows recorded twice, sessions whose doctor is not on file, a table with no key. The chapters run 223 queries against them, each printed with its actual result, and 64 figures show what the queries do.
The quickest way to run the examples is the LucidPlexus Workbench, a SQL notebook in your browser with the tables already loaded: nothing to install, and the data never leaves your machine. They also run unchanged in any copy of DuckDB. Data: lucidplexus.com/community/dataset.php?slug=healsync
For candidates preparing for a SQL interview, analysts a few months into the job, and spreadsheet and pandas users moving to SQL. You should be able to write a basic SELECT with a WHERE clause; everything beyond that is explained where it is first used.
Every purchase includes the companion download: the two-page cheat sheet, the workbook, the SQL file for every chapter with a script that checks each result against the book, and the six data tables.
LucidPlexus Premium. First edition, 2026.