All courses
100% free38 lessons available
Learn SQL
Learn SQL: query, filter, and sort data; insert, update, and delete rows; design tables with keys; join related data; group results; and use indexes and transactions.
38 course topics 38 in-depth lessons 0 topics planned
Start SQL course Learn SQLQuery · Model · Analyze
COURSE CONTENTS
Build the data model before the query gets complex.
This path is for beginners and working developers who want portable SQL they can explain before using PostgreSQL, MySQL, SQLite, ORMs, or reporting tools. Published lessons include runnable SQL, guided practice, and a knowledge quiz. The remaining course topics stay visible but do not link to unfinished pages.
01
Relational foundations
- SQL IntroductionSee how SQL, a DBMS, and the relational model fit together, then read a first table with SELECT.90 min
- SQL TablesRead a table as a contract: named columns, one meaning per row, and a schema that can change only on purpose.105 min
- SQL Data TypesPick a representation that matches the fact, then treat NULL as unknown instead of zero, false, or an empty string.120 min
- SQL Primary KeyGive every row a stable identity that other tables can reference, without treating a display name as the key.70 min
- SQL Foreign KeyStore a reference to a parent row, reject orphans, and model one-to-many or many-to-many links without lists in a cell.70 min
02
Reading and filtering data
- SQL SELECTRead a table with SELECT and FROM, project the columns you need, and name the output so another person can review it.110 min
- SQL WHEREKeep rows only when a condition is true, combine predicates with parentheses, and never compare a column with = NULL.125 min
- SQL ORDER BYMake result order a contract with ORDER BY, break ties with a unique key, and place missing values on purpose.55 min
- SQL LIMITReturn a bounded slice after ORDER BY. SQLite, PostgreSQL, and MySQL use LIMIT; other dialects may use FETCH FIRST.40 min
- SQL DISTINCTRemove repeated projected rows without changing stored data, and project only the columns whose duplicates should collapse.40 min
- SQL CASETurn conditions into readable result labels with CASE, match text with LIKE, and treat missing dates with IS NULL.130 min
03
Reports, aggregation, and analytics
- SQL Aggregate FunctionsCollapse many source rows into one measured result and know what COUNT(*), SUM, and AVG do with missing values.125 min
- SQL GROUP BYSet result grain with grouping keys, measure each group, then keep or drop groups with HAVING after aggregates exist.130 min
- SQL Window FunctionsKeep one row per source fact while measuring neighbors in a window, then choose a frame when a running total matters.135 min
- SQL CTEGive a complicated query reviewable names, then recurse only when a hierarchy needs a seed and an exit.135 min
04
Combining related data
- SQL JOINLine tables up through keys, keep only matching pairs with INNER JOIN, and check whether a one-to-many join multiplies rows.80 min
- SQL LEFT JOINPreserve the left-side population, treat NULL as no match, and avoid a WHERE that silently turns the outer join inner.70 min
- SQL Self JoinGive one table two roles with aliases, then walk a bridge table when many rows relate to many others.145 min
- SQL SubqueriesNest a query when it names the comparison clearly, and notice when correlation repeats work per outer row.145 min
- SQL UNIONStack queries that share a shape, choose whether duplicates remain, and remember UNION is not a relationship join.145 min
05
Designing reliable schemas
- SQL CREATE TABLETurn a requirement into a table whose name, columns, types, and first constraints you can explain.155 min
- SQL ConstraintsReject missing, duplicate, and illegal values at write time, and treat a constraint error as information.155 min
- Database NormalizationRemove update anomalies by storing each fact once, without treating maximum decomposition as a rule.155 min
- SQL ViewsGive a reusable query a name, hide a join behind a stable interface, and treat a stored snapshot as a different fact.155 min
06
Writing and protecting data
- SQL INSERTAppend rows with named columns so the mapping stays stable, and let defaults apply only when they are correct.55 min
- SQL UPDATEChange named columns on a reviewed population. A missing WHERE is a table-wide write, not a shortcut.50 min
- SQL DELETEForget a row only when it should not exist. A missing WHERE deletes the table's contents; a status flag keeps history.50 min
- SQL UpsertsHandle a uniqueness conflict in one statement. SQLite and PostgreSQL use ON CONFLICT; other engines may use MERGE.155 min
- SQL TransactionsMake a group of writes all succeed or all vanish. SQLite uses BEGIN; some engines also accept START TRANSACTION.155 min
- SQL IsolationTwo correct sessions can still empty a rule. Prefer database-side increments and version checks over remembered totals.155 min
07
Performance and administration
- SQL IndexesAdd an access path for real filters and joins, keep the column naked, and pay the extra cost on INSERT, UPDATE, and DELETE.155 min
- SQL EXPLAINA plan is the engine's recipe, not the SQL text. SQLite shows QUERY PLAN; PostgreSQL often uses EXPLAIN ANALYZE.155 min
- SQL Query TuningFix the question and the access path before adding indexes, and prove each change with a plan rather than a feeling.155 min
- SQL BackupsVersion the schema, restore parents before children, and compare counts instead of trusting a quiet INSERT.155 min
08
Security and production workflow
- SQL Users and RolesA session is who is connected. Least privilege is a small key ring; GRANT syntax is engine-specific and labeled here.155 min
- SQL InjectionValues are not source code. Bind parameters, parse integers yourself, and keep statement shape on the server.155 min
- SQL Stored ProceduresFunctions calculate, triggers react, and procedures bundle writes. Support and syntax vary widely by engine.155 min
- SQL Project: Production DataConnect tables, keys, joins, transactions, indexes, and injection-safe lookups in one production-minded catalog.155 min