New: Roadmaps ordered paths through our cheat sheets and flashcards, so you always know what to study next.
Explore themSee what's new on GitHubFrom relational tables to a PostgreSQL database you can run in production.
A 12-step learning path. Follow it in order, or jump to what you need.
For developers and data folks who want real command of SQL and relational databases, not just enough syntax to get by. Plan on about 6 to 8 weeks at a few hours a week, moving from table basics through schema design, performance, and a full PostgreSQL deployment. This path stays inside SQL and relational systems; if your work leans toward MongoDB and other NoSQL stores, the MongoDB & NoSQL Databases path covers that ground instead. By the end you can design a normalized schema, write and tune complex SQL, and run a PostgreSQL database with replication, security, and zero-downtime migrations in place.
Expected: basic comfort typing commands in a terminal or code editor. Helpful but not required: prior exposure to any programming language.
Open this first and you'll never again squint at a table without knowing what a key or a relation actually is, the mental model every SQL query in step 2 leans on.
Write real SELECT, JOIN, and window-function queries today, the vocabulary step 3's schema design and step 6's transaction commands both assume you already have.
Turn the tables you just queried into a schema that won't rot: normal forms, keys, and the anti-patterns that turn a clean database into a mess by month three.
You can write real SQL against a table and design a schema that won't rot as it grows. Next up: making those queries fast and safe under real traffic.
Finish this section to unlock.
+100 XP
Read an execution plan and know why a query is slow, the indexing skill that keeps the schema from step 3 fast once real traffic hits it.
Close the gaps that turn a SQL injection into a breach: encryption, roles, and auditing layered on top of the schema and queries you already built.
This is where concurrent writes stop being simple, ACID, isolation levels, and MVCC explain the deadlocks and phantom reads that step 2's plain BEGIN/COMMIT never warned you about, expect to revisit it.
You can index a slow query back to health and reason about isolation levels without flinching, the part most people bounce off (a few minutes of due flashcards keeps chapter 1's SQL sharp while you're here). Next up: putting all of it to work inside PostgreSQL itself.
Finish this section to unlock.
+100 XP
Go deep on the one RDBMS this path picks as its flagship: PostgreSQL's data types, planner, and admin tools turn everything from steps 2-6 into commands you actually run.
Keep a Postgres cluster alive when a primary dies: streaming and logical replication, failover, and the read-replica patterns that step 7's single-instance skills don't cover.
Take this if you're heading toward AI or RAG work: pgvector turns the PostgreSQL from step 7 into a vector store, so embeddings live next to the relational data you already trust.
Reach for this if you're shipping to a serverless or branch-per-PR workflow: Neon reshapes the PostgreSQL from step 7 into instant database branches and scale-to-zero compute.
Grab this if your job runs on MySQL instead of Postgres: the same DDL, joins, and isolation levels from steps 2 and 6, translated into InnoDB's syntax and quirks.
Ship a schema change without downtime: version-controlled migrations, expand-contract patterns, and CI/CD pull every skill from this path, design, SQL, security, and Postgres, into one deploy workflow.
You can design a clean schema, write and tune serious SQL, and run PostgreSQL in production with replication, security, and zero-downtime migrations in place, that's the SQL & Relational Databases badge, earned.
Finish this section to unlock.
+100 XP
Finish every required step, at least 70% of them genuinely done (not skipped), to earn this badge and 500 XP.